adventure-help

[sweet hornet]

Please provide an example of what you're trying to achieve, because it seems unclear

[frozen tree]

For example I have a component which is already parsed, how would I translate its papi placeholders now? (Without turning it back into a String)

[sweet hornet]

Why can't you parse the placeholders during the same time you parse your component?

[frozen tree]

Performance

[tawny bolt]

How would that make any performance difference, huh

[hallow tangle]

Components are immutable

[frozen tree]

I would imagine parsing 100 components each second take a little server performance

[tawny bolt]

nah

[frozen tree]

Nah or no? XD

[tawny bolt]

I do that once per tick for my scoreboard, no issues with 100 scoreboards

Parsing is fast

[frozen tree]

Mhh

Well, thanks for helping

[sweet hornet]

Did you actually encountered any performance issue?

[tawny bolt]

I recommend trying and profiling it for your usecase, and then see if its really an issue. Because that by far is the easiest way of doing the PAPI placeholders

[robust wharf]

That's a #paper-dev question

[fossil crest]

oh ok sorry

[ember quiver]

When trying to read raw playerdata from playerdata folder

File playerDataFile = new File(playerDataFolder, uuid + ".dat");

[prisma mason]

Packetevents is the failing plugin, not yours

[ember quiver]

hmm thats strange that packetevents brings kyori nbt

considering that ideally paper should've brought it

didn't notice from the first read, thank you

[obsidian nimbus]

well paper already has nbt builtin

[wet kestrel]

Are you sure you're running this on paper and not perhaps spigot or aaaancient paper?

[rare sage]

paper does not bundle adventure-nbt anyway

[carmine plinth]

I had the same issue. don't shade packetevents. use the snapshot version

the experimental jars

and plus being bad + adventirous shouldn't lead to a crash?????????????????

[sand drift]

wat

[carmine plinth]

yeah

being adventurous sometimes isn't safe. yeah be safe out there don't dig straight down

[rare sage]

they didn't shade PE lol, and that isn't a crash either

[sweet hornet]

Is there a component serializer for the new 1.21.5+ SNBT format?
I have a tool on 1.21.4 for my staff members allowing them to copy the components from an existing item's tooltip so they can use it when making other items, and I used the JSON component serializer because it was what Minecraft used in commands, but now we're upgrading to 1.21.7 and they need the SNBT encoding instead of JSON so they can still use in Minecraft commands

[cloud vapor]

What platform?

[sweet hornet]

Paper

[cloud vapor]

PaperAdventure has a codec somewhere

or is it AdventureCodecs

[sweet hornet]

Do I need something other than io.papermc.paper:paper-api:1.21.7-R0.1-SNAPSHOT ?

[sand drift]

it's an internal class

[sweet hornet]

Because I can't find such class (or at least, IntelliJ's autocompletion can't)

Ah

There is no way I can just serialize a Component into a SNBT String from the API? Sad

[sand drift]

SNBT is minecraft specific, and we don't expose SNBT inside of the paper API

[sweet hornet]

Hmm, and raw NBT bytes maybe? Serialization to SNBT is not an issue for me

[quaint monolith]

The bytes API is more comfortable because the API doesn't actually promise those bytes are in any particular format

https://jd.papermc.io/paper/1.21.5/org/bukkit/inventory/ItemStack.html#serializeAsBytes()

Doing anything with that byte[] other than saving it to provide back to the API unmodified is out of scope, if it breaks you get to keep both pieces

[sand drift]

there is some work on adventure for looking to expose a means of dealing with NBT, just, really unstable format

[sweet hornet]

Yes I know this for items, but is it accessible for raw components, or would I need to create a fake item with my component as display name so I can then extract it from NBT bytes after (which is a bit of a hack)

I don't really "care" about the unstable format, because I don't use this for storing. I just want it to be up-to-date format with the current in-game version, so that it can be used back in commands right after

[sand drift]

that does become a huge issue when writing a serialiser, however, which is the issue

if you're on paper your easiest option is going to be to use userdev and just poke internals for output

[sweet hornet]

Hmm ok I'll see how it turns out then, thx

[cloud vapor]

You can look at the nbt serializer PR

Haven't had a chance to look at it myself

[stable siren]

I'm updating my plugin to the newest adventure libs, and testing on newest Paper for 1.21.7

You are running the latest version

I've noticed messages to the action bar, and hover JSON components messages sent to the player are not functioning (no hover component when hovering over them) in my testing. I'm not seeing any stacktraces in console (maybe I need to enable logging levels? Would appreciate info on that if so.) nor were there any compiler errors or anything prior to testing. Just simply updated Paper for my test server, and updated the dependencies of all the adventure libraries I was using.

Any tips on how I can get to the bottom of what might be happening? When this has happened in the past I usually need to add some new dependency to my shadowed JAR that adventure has incorporated, but I would expect potentially an error or something to help identify that.

Thanks 🫡

[wet kestrel]

1. Are you using bukkit-platform or paper's API? I assume platform from your comment about shadowing.
2. Was your plugin working on 1.21.6? I know platform isn't even updated to that yet.

[stable siren]

I am using bukkit-platform
Testing on 1.21.6 also had the same issue, but I didn't mention it since 1.21.6 was around so briefly

[wet kestrel]

PR's here - https://github.com/KyoriPowered/adventure-platform/pull/212
It could really benefit from people testing it out and confirming it works.

-# Also just move to Paper API! :3

[stable siren]

I'll test it

@wet kestrel well I wanted to test, but the sonatype repo from md5 wants me to do some user/pass login stuff for some reason

   > Could not resolve net.md-5:bungeecord-chat:1.15-SNAPSHOT.
     Required by:
         project :adventure-platform-bukkit > com.destroystokyo.paper:paper-api:1.15.2-R0.1-SNAPSHOT:20201203.195914-323
      > Could not resolve net.md-5:bungeecord-chat:1.15-SNAPSHOT.
         > Unable to load Maven meta-data from https://oss.sonatype.org/content/repositories/snapshots/net/md-5/bungeecord-chat/1.15-SNAPSHOT/maven-metadata.xml.
            > Could not GET 'https://oss.sonatype.org/content/repositories/snapshots/net/md-5/bungeecord-chat/1.15-SNAPSHOT/maven-metadata.xml'. Received status code 401 from server: Unauthorized

I don't know the process to get an account with auth for that repo

actually that wasn't the only repo to reject me

[wet kestrel]

Oh, sonatype nuked snapshots repo didn't they

[rare sage]

tell that to fuuid

[wet kestrel]

I have two lines of code that exist for Spigot support for now. that's all 🙂

[tawny bolt]

if (spigot) {
 dont(); }

[wet kestrel]

New snapshots repo if you're up for editing more than you thought you would: #build-tooling-help message

[rare sage]

are you telling me fuuid exclusively supports modern versions of paper?!?!

[wet kestrel]

No, and let's stop misusing a help channel :3

[rare sage]

[stable siren]

Looks like I've run into some unfortunate timing

after the build is updated for the new repos, I don't mind testing that PR
I changed the build locally, but it's not finding the artifacts, I suspect I made a mistake in my changes and it's probably my fault, or perhaps these artifacts are somehow not available yet (maybe unfortunate timing from today being the day it got nuked)

[sweet meadow]

Just want to ask before i spend hours trying to do this to realise its not possible
Is it possible to use adventure nbt to (de)serialize things like block states (or block data, whatever is needed to copy the exact data of the block)
if so, is it platform agnostic (will it work with serializing paper block state and deserializing as a fabric block state)

if the answer to all that is yes,
are there any libraries that do this or am i going to make my own serializers

[cloud vapor]

NBT is just NBT, if you've got NBT, adventure-nbt can handle it

[sweet meadow]

yeah worded this badly sorry, is there built in methods to convert native bukkit objects to nbt

[sweet hornet]

Adventure does not know about any bukkit stuff

[sweet meadow]

paper knows about adventure though and there is converters for text related things

[vocal dove]

Is there a way to convert an Adventure component to an NMS component?

[midnight spire]

that's something that would be in adventure-platform, not sure if it's exposed in some form of API though

I kinda doubt it tbh

ok apparently MinecraftComponentSerializer exists and seems accessible

[rare sage]

it depends on the platform yes

not all platforms have a MinecraftCS

e.g. -platform-modded has asAdventure/asNative methods in the MinecraftAudiences type, -platform-bukkit has MCS, native platforms will have their own ways too

[mint path]

How do you properly parse unicode glyphs? I use something like this atm: java if (unicode != null && unicode.contains("uE")) { unicode = StringEscapeUtils.unescapeJava("\\" + unicode); but that doesn't work with Adventure. /tellraw works but my usage of some unicode chars becomes empty content and doesn't work.

[cloud vapor]

what do you mean by "parse"

[mint path]

well they come in like this

[cloud vapor]

sounds like your file encoding is borked

[mint path]

nvm i had some unicode strings with a lowercase e so that .contains("uE") wouldn't fire on em

[cloud vapor]

what im saying is you don't have to do any parsing or unescaping logic if you just have your file encoding set correctly

[oak hamlet]

Hey! I've got weird error when I am trying to start my server with custom plugin:https://pastebin.com/KjEU1ji8
this is my pom:

I've had the same error on the other project that was built with gradle.kts and relocating adventure solved it. The version is the same - 1.17.1 but the error still stays

[cloud vapor]

You cant relocate adventure and use native paper component methods

java.lang.NoSuchMethodError: 'void org.bukkit.inventory.meta.ItemMeta.displayName(pl.karoldronia.sponges.libs.adventure.text.Component)'

See how it's looking for this method which definitely doesn't exist

[oak hamlet]

yea I see

but what should I do to fix it?

[cloud vapor]

This https://docs.advntr.dev/faq.html#why-am-i-getting-a-nosuchfielderror-nosuchmethoderror-classnotfoundexception-or-similar-when-updating-using-adventure-platform-adventure-text-minimessage-adventure-api-or-other-related-libraries-tools

[wet kestrel]

tl:dr stop shading and relocating adventure. It's in paper.

[oak hamlet]

okay I don't know if I do it correctly but this is my pom:

       <dependency>
            <groupId>net.kyori</groupId>
            <artifactId>adventure-platform-bukkit</artifactId>
            <version>4.3.1</version>
        </dependency>

        <dependency>
            <groupId>net.kyori</groupId>
            <artifactId>adventure-text-minimessage</artifactId>
            <version>4.18.0</version>
        </dependency>

I am not including adventure-api and I am not relocating anything
Error:

public interface MiniMessageHolder {

    MiniMessage MINI_MESSAGE = MiniMessage.miniMessage();

}

[wet kestrel]

I can't open attachments easily on mobile

Also share your full build script

[oak hamlet]

gimme a moment

https://mclo.gs/R962aOh error

https://mclo.gs/tqqkPqG

pom

[cloud vapor]

paper already ships all that so idk why you're including it anyway

and if you're not using paper, then dont use paper

[oak hamlet]

so jsut remove all adventure dependencies?

[wet kestrel]

I suspect that you're building against a much newer adventure than the one bundled in the ANCIENT (outdated, unsupported, exploitable) version of paper you're using.

[oak hamlet]

org.bukkit.plugin.InvalidPluginException: java.lang.NoClassDefFoundError: net/kyori/adventure/platform/AudienceProvider

[cloud vapor]

Paper includes adventure - Player is an audience, you don't need an AudienceProvider

[wet kestrel]

I feel like you haven't removed your adventure deps if you're even able to build with that...

[oak hamlet]

yea now I have removed all adventure dependencies but I got:
java.lang.NoClassDefFoundError: net/kyori/adventure/text/minimessage/MiniMessage do I need to implement MM?

[obsidian nimbus]

what version you running?

[oak hamlet]

server is on 1.17,1

[wet kestrel]

You're in such ancient territory there. Not sure anybody remembers specifics. You should upgrade off a server full of issues to modern.

[obsidian nimbus]

Either upgrade to the latest paper version, or either use adventure-platform with relocation. (you will lose ability to use native methods)

[cloud vapor]

Yeah we can't provide support for legacy paper builds

[frank glacier]

any forecast to support 1.21.7?

[deep grotto]

There is no ETA

[wet kestrel]

I didn't think there were changes between 1.21.6 and .7 that were adventure-relevant. Are you perhaps using adventure-platform-bukkit?

[frank glacier]

yes

4.4.0

[wet kestrel]

Okay, then you want #adventure-platform-mod-help where you can see conversation about testing the PR adding that support. Adventure itself has been fine for a whiiile.

[scenic cloak]

HELP, server on 1.21.7
player.sendMessage(
Component.text("test")
.hoverEvent(HoverEvent.showText(Component.text("here")))
.clickEvent(ClickEvent.runCommand("/say test"))
)
clickEvent doesn't work but other events work

[cloud vapor]

What platform?

[sterile rampart]

isn't that just a signed argument thing?

[wispy wagon]

how can I serialize a component as adventure nbt?

[molten berry]

https://github.com/KyoriPowered/adventure/pull/1084

[wispy wagon]

Yeah, I figured it out myself now. But I would assume its not in any repo yet and I'm too lazy to setup that myself so I just added this into my codebase lmao

[molten berry]

I didn't know Kotlin had async await

[wispy wagon]

Too lazy to write "Wait for" lmao

[wispy wagon]

I mean technically you can do .await() on a completable future in a coroutine

[gritty tundra]

can I send a translationkey or translatable component as a moonshine @Placeholder or do I need to make a custom resolver for that

[wicked torrent]

iirc moonshine does not integrate directly with adventure at all

so it's kind of up to you to choose how to fit them together

but you can definitely do that

[obsidian nimbus]

isnt moonshine dead

[gritty tundra]

pretty much

[fossil prawn]

What's the best way of detecting whether an audience can use clickable components or not?

Which is typically just console

[quaint monolith]

Console doesn't work with clicks? I guess I could see the URL part being hard/impossible but run and suggest command should be doable

That's doing to be dependent on the platform implementation though and I don't think there is any way for you to know

[fossil prawn]

Are certain terminals meant to work with clickable components?

[quaint monolith]

They could, I don't know if any of them actually do work though

[fossil prawn]

hmm ok. Thank you

[quaint monolith]

Whether or not console can click on things is going to depend on the kind of click event, the platform, and what terminal you're running the game in

[cloud vapor]

why do you need to detect it?

[quaint monolith]

Heck even URL click events would work if they just called out to xdg-open

[smoky wadi]

Does minimessage support shadow gradients as of now? I saw a github pull request.

Like this

[sweet hornet]

If the PR isn't merged, then no

You could technically do it yourself, or up-vote the PR and share your thoughts about it to help the team

And, as we can read in the description:

Done as a quick tech demo, would need a bit of actual design work and cleanup if we want to actually incorporate this into MM
It's not gonna be merged like this, so you should rather leave a comment explaining why you consider such feature to be useful, and worth investing some time to integrate it properly

[robust wharf]

Can also copy and adapt the tag from the PR into your plugin

[quaint monolith]

Huh, that's way better than gradient on the text itself

[green monolith]

Yeah, that looks really pretty

[sweet hornet]

iirc when I tested it, it looked nice on white text with (multi-)colored shadow, but any text color other than white was not that good in terms of results

[merry valve]

Hello, how can I reset text decoration like &r in Component.text(...?

[sweet hornet]

.decoration(TextDecoration.ITALIC, false)

[merry valve]

so that disables italic? do I understand right?

[sweet hornet]

That's it, if you need to remove other decorations, you can just replace ITALIC by something else

[merry valve]

Thanks!

[tough peak]

What does MiniMessage.miniMessage().serialize(component); do? please provide an example

[robust wharf]

what do you think it does? lol

it turns a component into a minimessage string

[tough peak]

ok thank you

[obsidian nimbus]

lmao

[livid blade]

So I do not know if im blind, but im trying to find the docs to do with the click events for adventure?

I checked the adventure docs quite a few times and feel daft for not finding the right page

[wet kestrel]

https://jd.advntr.dev/api/4.21.0/net/kyori/adventure/text/event/ClickEvent.html ?

[livid blade]

No actual docs? Just the javadocs? Thats a shame, but cheers 🙂

[wet kestrel]

The docs describe there are events. The javadocs show how to use them. They're not super complicated.

If you feel there could be an improvement in the docs, feel free to write an issue or pull request.

[frozen tree]

Hey, why does the LegacyComponentSerializer not preserve the reset tags(&r and <reset>) upon deserializing a legacy string.

My code:
miniMessage.serialize(legacySerializer.deserialize(msg))

[wet kestrel]

Those two things (reset tag and legacy r) are not identical. It makes more sense to deserialize the legacy to something that makes sense as a tree in components, and then minimessage serializes the tree.

-# Also <reset> is not something you should encourage others to do often :3

[livid blade]

Where? Genuinely didn't see them haha

[stable siren]

@wet kestrel I was able to compile the build now for blood's PR for platform-bukkit and I was able to confirm it working in Paper 1.21.7 builds, I left a comment on the PR. Sorry I was unable to test it the other day (unfortunate timing with that snapshots repo being sunset... or nuked or whatever)

[wet kestrel]

https://docs.advntr.dev/text.html#events mentioned here

[wet kestrel]

Huzzah! Thanks for helping! 🙂

[frozen tree]

Isn't that exactly what my code is doing? It deserializes the legacy string into a MM object using the Legacy Serializer, and then MM serializes it back into a string

[wet kestrel]

Yes. I am describing what your code is doing. It deserializes the legacy stuff into a tree. The tree lacks the concept of "reset". Then it serializes the tree as minimessage.

[livid blade]

Oh wow my bad, cannot believe I didnt see that as I even used the builder example earlier to change my Components due to the immutability.

My apologies once again

Also semi neat / odd thing question... do you use dark mode on the docs by chance mbax?

[wet kestrel]

I do not. Never bothered to hit the button. My OS defaults to light.

[frozen tree]

Right, got it, that might be my language barrier 😅 Either way, so isn't there any other way to make it deserialize the &r aswell?

[wet kestrel]

I do not think so. And that is a bad design to do so.

[frozen tree]

Why? I mean, how would one reset for example strikethrough without it?

[fiery oracle]

components have no concept of "resetting"
the &r in legacy text basically just means "wrap the following text with an explicitly set style (white text, no formatting or events)"

this is also notably different than the MM <reset> tag, which closes currently open tags

[wet kestrel]

Strikethrough is a different code from r and should process fine.

[frozen tree]

Okay, well, thank you for the help

[rustic pumice]

What does component.color(); return if multiple NamedTextColors are used in the component?

[rare sage]

a component can only have one colour

however, a component can have children which have different colours

Component#thing() will return you the thing that one component in the tree directly holds

[smoky girder]

you could analyze all the colors in use using the ComponentFlattener

[rustic pumice]

Im guessing .append() makes those component children? also .thing() doesn't seem to exist at all, looked through javadocs

I feel like this is really dumb, but is it possible to get a List of (or at least separate) components that have .append() in it?
What im trying to do is get the colour of the player's name from displayName() in Paper, with other text being in the display name

[rare sage]

.thing() was an example that applies to any one property of a component

children, content, style or style elements like colour or decorations etc

.thing() is not literally an actual property of a component

i'm not entirely sure what is it you're asking, components are a tree, you can traverse them with the various iterator/iterable methods on Component, or with a ComponentFlattener

[slate current]

Would a PR to adventure NBT adding something like

List<MyData> values = ...;
ListBinaryTag.Builder builder = ListBinaryTag.heterogeneousListBinaryTag().initialSize(data.size());
for (MyData data : values) {
    builder.add(StringTag.stringTag(data.thing());
}

be fine? basically just, a way to initialize builders with a set size for when you're directly copying another list/map into it.

(Or is that already a thing somewhere and I missed it?)

[cloud vapor]

Do you have a particular reason for wanting this?

[slate current]

Just some potentially hot code that I'd like to optimize all the way™, I know it's not the sort of thing that'd make a massive difference, but any easy & non-messy optimization is a good optimization

[sage river]

if you already want to do something like that then I suppose it would be better as heterogeneousListBinaryTag method parameter, so you don't create the backing list object twice

[slate current]

Also unrelated question - is there a reason a lot of defaultValue parameters are @NotNull? compound.getCompound("key") returns an empty compound if the key is missing, so I was gonna do .getCompound("key", null), but ^

[slate current]

Wouldn't be a problem currently as builders only initialize their underlying list/map when something is first added as far as I can see, but if they want to treat that as a subject to change implementation detail thing then yeah probably in the creation method

[cloud vapor]

does the code enforce non-null on the param?

[sand drift]

Nope

[slate current]

It doesn't enforce it, but like, it's an IDE warning

[cloud vapor]

then yea just an oversight, happy for the param to be nullable with a contract annotation on the return types

[slate current]

I suppose you could do

List<BinaryTag> tags = new ArrayList<>(data.size());
for (MyData data : values) {
    tags.add(StringTag.stringTag(data.thing());
}
return ListBinaryTag.listBinaryTag(BinaryTagTypes.LIST_WILDCARD, tags);

For mostly the same effect?

Still has some redundant list copies, but that is probably just micro optimization at that point (and I assume you wouldn't want a fromUntrusted(List) either way)

[sweet hornet]

Just out of curiosity, what is approximately the size of your lists you're talking about?

[slate current]

Any size, which is why I preferred to do it like this - could be very big, could be very small

[cloud vapor]

yeah just stick a param on the builder methods

would be fine for both

[slate current]

Alright, thanks!

[slate current]

Sorry, haven't used @Contract much - is @Contract("_, null -> _; _, !null -> !null") + making the method itself nullable correct for that?

If the second param is null it may return null, if the second param isn't null it will never return null

[sweet hornet]

iirc the first part is useless, because it does not specify anything about the return value

[slate current]

Ah, makes sense

[sweet hornet]

Only _, !null -> !null is needed to say that, whatever is the first value, if the second one is not null, then the return value won't be null

[slate current]

Yeah I see, thanks - it already knows it might be null because it's nullable, just need the contract to tell it when it isn't

[sweet hornet]

In this case, yes. But it could also be the other way around, in some other cases: _, null -> null for example, says that the result will always be null if the second param is null. Is says nothing about when it won't be null, meaning it can still be null even if the second param isn't. In this case obviously this is not true, but I was just showing you an example. The method could be "get the property of that user, which might exist or not, but will always return null if the user is null"

[cloud vapor]

you need _, null -> null; _, !null -> !null

[sweet hornet]

Wait, no? The result won't always be null when the default value is: There is the case where the requested value exists and is returned

Unless I'm missing something, I just gave a really far look at the problem

[rare sage]

_, null -> null doesn't make any sense

you're saying it will always return null if the second param is null

it's just _, !null -> !null

[cloud vapor]

that's not what contract does, null just indicates the return type is nullable

[sweet hornet]

No, I really don't think so, the nullability of the return type is indicated with the @Nullable on the method

[rare sage]

https://javadoc.io/static/org.jetbrains/annotations/26.0.2/org/jetbrains/annotations/Contract.html
that just means "ignoring what the first parameter is, the second parameter being null results in a null return value", not a null return type

[forest meteor]

when will adventure be ready for 1.21.7? im having a bug with the gson serializer and im not sure how im going to handle it im ngl

https://paste.gravemc.net/?id=sfHC0LWQ9s

here is the error im having

public static final GsonComponentSerializer GSON_SERIALIZER = GsonComponentSerializer.gson();```

literally whats failing rn

[slate current]

Is there any reason you're shading adventure? iirc the GSON serializer is included in Paper

[rare sage]

you are not relocating net.kyori.option

[forest meteor]

lots

[forest meteor]

oh tysm

[slate current]

PRed these two, but I have another question - is there any way to get a builder from a compound tag (other then making an empty builder and adding the values from the tag)?
The put methods on CompoundBinaryTag copy it and return a new one (which is very useful actually), but then when editing multiple values at once it'd be creating a copy every time, which isn't great.
So something like

CompoundBinaryTag edited = tag.createBuilder().put(x, y).putInt(x, y).remove(x).build();

that'll create a builder directly from the compound's map could be very useful (if it doesn't already exist)

[round fox]

Is there any version of a MiniMessage parser that's able to be ran outside of Minecraft? Like on a non-minecraft app

[sand drift]

I mean, MM is tied to adventure itself, it's just a serialiser

it's the component handling which is where platform stuff starts to come into play

[round fox]

I see

I know there's a minimessage web viewer so I thought maybe there was a 'generic' implementation somewhere

[sand drift]

Well, yea, that's just MM + Adventure

[cold junco]

I have
<click:run_command:/kick <player>> and replace <player> using tag resolver. <player> is parsing, but <click> is not.

btw if I do <click:run_command:'/kick <player>'> <click> is parsing, but <player> is not

how to fix it?

[wet kestrel]

What kind of resolver is player? I make sure to use placeholder.parsed, so I don't mess it up.

[cold junco]

it's placeholder.unparsed

hm, it worked with parsed. thanks

[wet kestrel]

Yep! 🙂

[livid blade]

Is there a way to use the ClickEvent of a component to enact a lambda function / method on click? If so how? If not, is there a reason why that was never added? It seems like a reasonable thing for it to do considering how Paper-entire is based around methods called on event based logic 😅

e.g: imagine having a class defining some custom function, and you can set the click event to call a method within that class if the component is clicked?

[sand drift]

ClickEvent.callback or something

[livid blade]

See this is why I politely suggested maybe the docs need a tiny bit of elaboration / a few examples of ClickEvent being used in various use-cases.

As I just looked it up and it says the param is a ClickCallback<Audience> which is taking me further down the "undocumented (other than javadocs) abstractions" rabbit hole 😅

[sand drift]

I mean, the docs always need more love, be the change you want to see

[livid blade]

*I would if I knew what to change.

Hope you see the issue here haha

[sand drift]

Pretty sure ClickCallback is just a functional interface

[livid blade]

There is a certain level of "Ive figured it out" fatigue.

By the time you brute force it or bash head up wall till it works, you usually dont have the energy to go fix the docs.

Sometimes we just need someone a bit more knowledgeable to come in and add an example for us poor unfortunate sods haha

[sand drift]

Yea, can basically just ClickEvent.callback((audience) -> { audience.sendMessage(text("boo"));};

[livid blade]

On a slight tangent, the Tauri documentation is rife with "Ive figured it out fatigue".

So much effort goes into getting stuff to work, theres no energy left to fix the docs haha

Cheers Electro

[livid blade]

/**
 * Represents a collection of messages that can be shown in bulk to an Audience.
 */
public class MessagePage {

    /**
     * The title of the page.
     */
    private final Component title;

    /**
     * The main content of the page.
     */
    private final List<Component> content = new ArrayList<>();

    /**
     * The footer (the final message at the bottom of the page).
     */
    private Component footer = Component.empty();

    public MessagePage(Component title, Component footer) {
        this.title = title;
    }



    /**
     * Adds a new component to the content of the page (placed directly below the previous line of content).
     * @param content The component to add.
     */
    public void addContent(Component content) {
        this.content.add(content);
    }

    /**
     * Shows the page messages, in the correct order, to the specified Audience.
     * @param audience The audience to show the page to.
     */
    public void showPage(Audience audience) {
        TextComponent.Builder builder = Component
                .text()
                .append(title)
                .appendNewline();

        audience
                .sendMessage(addContent(builder)
                        .append(footer)
                        .build()
                );
    }

    public void setFooter(Component footer) {
        this.footer = footer;
    }

    private TextComponent.Builder addContent(TextComponent.Builder builder) {
        if(content.isEmpty()) return builder;
        for (Component component : content) {
            builder.append(component).appendNewline();
        }
        return builder;
    }
}

public class MessageStack {
    private final List<MessagePage> pages = new ArrayList<>();
    private int currentPage = 0;

    public MessageStack() {}

    public boolean addPage(MessagePage page) {
        return pages.add(page);
    }

    public boolean addPage(Component title, List<Component> contentLines) {
        MessagePage page = new MessagePage(title);
        for (Component content : contentLines) {
            page.addContent(content);
        }
        return pages.add(page);
    }

    public void showCurrentPage(Audience audience) {
        if (pages.isEmpty()) return;

        MessagePage page = pages.get(currentPage);
        page.setFooter(generateFooter());
        page.showPage(audience);
    }

    public void nextPage(Audience audience) {
        if (currentPage < pages.size() - 1) {
            currentPage++;
        }
        showCurrentPage(audience);
    }

    public void previousPage(Audience audience) {
        if (currentPage > 0) {
            currentPage--;
        }
        showCurrentPage(audience);
    }

    private Component generateFooter() {
        TextComponent.Builder builder = Component.text();

        Component pageLabel = Component.text("Page " + (currentPage + 1) + " of " + pages.size())
                .color(NamedTextColor.GRAY);
        builder.append(pageLabel).append(Component.text("   "));

        if (currentPage > 0) {
            Component prev = Component.text("[Previous]")
                    .color(NamedTextColor.YELLOW)
                    .decorate(TextDecoration.UNDERLINED)
                    .hoverEvent(HoverEvent.showText(Component.text("Go to previous page")))
                    .clickEvent(ClickEvent.callback(this::previousPage));
            builder.append(prev).append(Component.text(" "));
        }

        if (currentPage < pages.size() - 1) {
            Component next = Component.text("[Next]")
                    .color(NamedTextColor.GREEN)
                    .decorate(TextDecoration.UNDERLINED)
                    .hoverEvent(HoverEvent.showText(Component.text("Go to next page")))
                    .clickEvent(ClickEvent.callback(this::nextPage));
            builder.append(next);
        }

        return builder.build();
    }

}

Im pretty happy with that as an MVP (will be changing the formatting of the page buttons obviously, but the concept is pretty neat) 🙂

Nice little collection of messages / logs / whatever components you fancy into a "page" of messages you can flick between nice and easy.

Wish there was a slightly more performant way to do page footers than re-do them constantly every time you flick the page.

Maybe builder pattern for the MessageStack, so when the footers are added there is 0 chance of new pages being added?

[fiery oracle]

is that bottom most private method in the 1st one supposed to still be there or

oh and the first one could implement ComponentLike instead of having a method to show to an audience

same with the 2nd one if the page moving is separated out from the showing

[wet kestrel]

But it's right there in the javadocs? Second method in there, too. Javadocs should be your first place for "what can I do with [classnamehere]?"

[livid blade]

Why would I want to not show the audience via the class itself?

[livid blade]

I appreciate that, but, in my defence, I'm very stupid 😉

Jokes aside it couldn't hurt.

A good example is i have no idea what ComponentLike is, so im gonna pop both to the docs and javadocs haha

[sweet hornet]

It would allow you to instead of msgPage.showPage(audience) you would do audience.sendMessage(msgPage). The main difference being that you could easily replace sendMessage by any other method, and also merge it with other components or anything.
An ComponentLike is just something that can be used as a component (thanks to the method asComponent)

[livid blade]

What other commands would be useful to have do to a page other than send Message?

[sweet hornet]

Idk you exact use-case, but I could imagine showing it through a book, a dialog, or doing other kind of component-level operation such as appending a prefix / suffix, or replacing texts, colors, styles…

[raven tusk]

Can I get translated item name as String using minimessage?

[sweet hornet]

What do you mean? What are you trying to get? The plain text of the name? The MiniMessage-encoded component? Something else?

[raven tusk]

Translated Item name as String.java from tranlation component

[sand drift]

MM just deals with serialisation

if you wanted to have the localised names in there, you'd need to flatten them

[sweet hornet]

Furthermore, "translated" in which language? English? Korean? iirc on Paper they bundle the English locale file, but in other case you'd need to include and maintain the translation file as well if you want to do server-side translation

[raven tusk]

So, is there no way to use a Translation Component for item names with Bukkit/Adventure in Spigot either?

[sand drift]

that stuff will work as much as spigot allows it to work

adventure mostly covers UX related stuff, it's just a conveinent library that we're able to adopt into the paper API across the board; I don't think that there is a public adapter for non UX stuff;

[sweet hornet]

A translation component contains a translation key with translation arguments (the values for the "placeholders" inside translated content), as well as the style (color, decoration...). You can very much use it, but it won't give you the translated content alone: you need to use a translator, fed with the translation file of the language you want to translate into

But it seems that what you want to achieve is unclear, even maybe for yourself. Please explain a little bit more (with context) about your use-case so we can better understand your goal to actually hepl you achieve it

[sand drift]

Yea, I was gonna say, assuming I even understand what you're asking; But, adventure platform doesn't deal with items; idk if they come with a flattener preconfigured to be able to use mojangs built in localisation file for english either

[slate current]

And two more things actually -

1. would you be open to adding CompoundTag#containsKey(String) and CompoundTag#contains(Key, Type) utility methods?
2. is there any way to get a BinaryTagType by its byte id? I can see an internal method, but as far as I can see it isn't exposed anywhere - would be nice if that could be made public, or at least a way to get a list of all BinaryTagTypes so that I could build my own mapping.

Sorry for all of the NBT questions btw lol - migrating a pretty big project from an old NBT lib to adventure NBT, so I'm just trying to PR some stuff that I'm missing

[daring wyvern]

Hello yall i got a question im tryna learn plugin making with paper on kotlin but i cant quite figure out whats the correct return type after i call ClickEvent.callback(foo()) for foo()

for context this is the foo function im trying to call

[robust wharf]

Nothing

Wait that looks strange

Please share more of your code, what is that function supposed to do and why do you want it to return a click callback

[daring wyvern]

this is more of a test to see if i can understand how it works so its veeery ugly

but my main goal is to fire handleClick when i click the text

my main problem is it excpects a return value and if i make it nullable it errors out on the callback

[sterile rampart]

ClickCallback is the functional interface type

you're invoking your callback when creating the click event. you should be passing that callback function in instead

[robust wharf]

So like, blah.clickEvent { handleClick(ctx) }

[daring wyvern]

trying this but it complains abt type mismatch

[rare sage]

that's just pseudocode example

look at the method in the interface, that's the signature your lambda has to look like

[daring wyvern]

bruh im so dumb lmao

[robust wharf]

Sorry would be blah.clickEvent(ClickEvent.callback ({ handleClick()}))

Or does kotlin let you extract the lambda from there? Idk, kotlin weird

[daring wyvern]

im too used to c# lambdas im tryna wrap my head around kotlin syntax

very clean but my small brain cant keep up

[robust wharf]

It's not super beginner friendly

It was written by senior java devs to speed up writing code I guess

[slate current]

No worries if you're busy right now, but is there a chance you could quickly 👍/👎 my suggestion above, just so that I know whether I'm good to PR it?

[daring wyvern]

ok so i fixed it with this line, as far as i understand clickEvent wants an event to summon when its time to invoke it and the callback then simply acts as a "pointer" to what to run (i may be extremely wrong)

thx for yall help anyways

[cloud vapor]

First sounds good, idk about the second

but if it's not too much hassle just PR it and we'll see

[slate current]

Will do, thanks!
And how about the earlier one (#adventure-help message)?

[slate current]

None of these are a ton of work code wise btw, so if you'd rather I just PR them/any other small things I may have in the future so that everyone can take a look, let me know

[cloud vapor]

yeah that also sounds good to me

[slate current]

Is there a cleaner way to do this?

ListBinaryTag listTag = compoundTag.getList(key, BinaryTagTypes.STRING);
for (BinaryTag tag : listTag) {
    materials.add(((StringBinaryTag) tag).value());
}

(As in, without having to cast the tags, since we know the list is of that type)

[prisma mason]

Is this possible?
<start_date:'<date_format:'dd.MM.yyyy':'EEEE'>'>

first one being resolved by a date resolver and the second being a booleanChoice

[wet kestrel]

Why not just have your statt_date take in the ddMMyyy stuff and not have it run through two resolvers?

[prisma mason]

can you elaborate? I thought of one way by just using Java's DayOfWeek's display name (PlaceHolder.unparsed usage)

[wet kestrel]

I'm saying to just do <start_date:'dd.MM.yy':'EEEE'>

[prisma mason]

how would that work?

[wet kestrel]

What do you mean? No differently from any other tag resolver you make.

[prisma mason]

what...

[wet kestrel]

Can you clarify what you're confused about? 🙂

[cloud bison]

How would I use a custom click event type via Minimessage? I get something like <click:custom:...> but not sure how to format the "..." since the payload is more complex

[smoky girder]

based on the javadocs, the payload is a key and an NBT tag

honestly i don't understand much more than that, if the point is to run code when the text is clicked then classically run_command is used

trying to figure out where the server receives the custom payload and i dont see anything

i guess this packet?
https://mappings.dev/1.21.7/net/minecraft/network/protocol/common/ServerboundCustomClickActionPacket.html

ah! here we go

this is the corresponding Paper event
https://jd.papermc.io/paper/1.21.7/io/papermc/paper/event/player/PlayerCustomClickEvent.html

[cloud bison]

Yeah, appreciate it, though I know about the event and everything. The reason for the custom action is so that you can run custom code without having to register a custom command to do so

Just trying to see how to do it through minimessage, I've tried the things that make sense to me but can't get it working

[smoky girder]

the click tag seems to be hardcoded to only accept 1 argument after the click type
https://github.com/KyoriPowered/adventure/blob/2f48b6b0cd9332781f7ec724012391f4a70c663b/text-minimessage/src/main/java/net/kyori/adventure/text/minimessage/tag/standard/ClickTag.java#L66

so its not like you can pass the key and then the NBT

[cloud bison]

I believe the NBT is optional for custom events

But then I'd think something like this would work "<click:custom:'minecraft:test'>Hey</click>", but it doesn't seem to

[fiery oracle]

the minimessage tag requires a string value
it doesn't work at all with the custom type because the tag only operates on strings

it using a string payload:
https://github.com/KyoriPowered/adventure/blob/main/4/text-minimessage/src/main/java/net/kyori/adventure/text/minimessage/tag/standard/ClickTag.java#L66
it not supporting non-text payloads:
https://github.com/KyoriPowered/adventure/blob/main/4/api/src/main/java/net/kyori/adventure/text/event/ClickEvent.java#L248

[smoky girder]

you can always write your own tag resolver that works in the way you'd like

it could even prepend a namespace of your choice, to avoid the escaping issue

so <mycustomtag:foo> would create a custom click event with key mynamespace:foo and empty NBT

[cloud bison]

I guess I'm just surprised because hover:show_item also needs to accept a key, which it accepts as a string, but doesn't work for custom whose only required field is id which is also a namespaced ID

But that makes sense, custom resolver seems like a good approach so I can do that, cheers gang

[cloud vapor]

is the custom click event even implemented in minimessage

I don't think I did it

[slate current]

With BinaryTagIO, is there any way to unwrap lists as you load them?

Kinda like NMS loading logic does:

(And if not, would that get accepted as a PR? otherwise you'd have to have loop over every tag in every tag to make sure they're all unwrapped as far as I can see)

[slate current]

Implemented something like that ^ for now, but would be a lot nicer if the deserialization could do that automatically (probably with some param to toggle it)

[hoary plume]

Question:
Would there be a way to detect if a specific feature is actually available on the server/proxy Kyori is being used on?
Like in my case would I like to know, if a Server/Proxy runs a version supporting Dialogs to use it.

Also, is there any form of tutorial or example on how to use Dialogs in Adventure?
I want to make a platform-agnostic Dialog system, so that I only have one core code base for it that works on proxy and server (In my case would it run on either Paper, Velocity or BungeeCord (using the platform library), tho I consider droping the last one).

[sweet hornet]

Adventure does not yet provides dialogs

Some platforms do (Paper)

[slate current]

^ I think there's only DialogLike stuff for platforms to extend, but not a full API yet

[hoary plume]

Rip

[sweet hornet]

And it's not likely to happen soon

It will, at the end, be the case. But when it will be stable

[hoary plume]

Guess I'll have to make platform-specific code then to create the Dialogs...

[sweet hornet]

For now, you have to use Paper's implementation if you want to use them, and idk about other platforms for now

[wet kestrel]

Dialogs are too potentially in flux. Paper's implementation has the whole package marked experimental.

[hoary plume]

And this?
Is there a way to know if a Dialog(Like) would work?

I assume I would need to do Class.forName(...) checks for the respective platform?

[sweet hornet]

Are you talking about detecting the presence of an implementation (depending on the platform), or the version of the platform (Paper...) being < 1.21.7 (or whatever version dialogs first came out)?

[hoary plume]

I simply want to make sure that when I want to use Dialogs, that the platform the server runs on is actually a version that could support it.

[sweet hornet]

Well, since implementations are provided by platforms, if the platform provides one, then it certainly exists, otherwise no. If you're talking about backward compatibility, then you need to check in which version of each platform the dialogs implementation came out, and simply test that version (or, if you don't want to bother checking version, you could technically test for the presence of the specific implementation class, yes)

But you cannot reliably platform-agnostically determine whether dialog implementation exist or not, since dialog implementations are provided by platforms. The best way you can do is assume there's not, and for specific platform you know there's one, then use it

[prisma mason]

I know I've asked this before, but can I make nested tags in a choice tag or should I just use a custom tag resolver?

For the datetime thing I mentioned before I used a custom tag resolver

[forest meteor]

yo, after updating adventure click event doesn't work for older versions

idk if thats because of Adventure or Adventure Platform

but what i know is that it works for 1.21.7 but not older versions

[wet kestrel]

If using -platform use #adventure-platform-mod-help 🙂

[forest meteor]

if i have a decoration/style that is set to false rather than unset.. how do i make it unset

[prisma mason]

https://jd.advntr.dev/api/4.23.0/net/kyori/adventure/text/Component.html#decoration(net.kyori.adventure.text.format.TextDecoration,net.kyori.adventure.text.format.TextDecoration.State)
with this

set the state to NOT_SET

[hoary plume]

Is there any form of info/guide for how I could create some own Dialog implementation using Adventure?
Would like to implement some basic system for Velocity to display Dialogs to a player.

[cloud vapor]

You would need the platform to implement it - adventure only provides the framework for platform dialog API until we've added a dialog api of our own

[hoary plume]

Guess I have to wait until Velocity has something...

[sand drift]

The issue I have with velocity is, well, time, but, also; Do I implement something that I will likely then need to deal with again in the future to make it conform to the adventure API

[small cloud]

is it possible to make a resource bundle from a file instead of a in jar resource?
https://docs.papermc.io/paper/dev/component-api/i18n/#resourcebundle

[cloud vapor]

Yes, there's methods on the string based translation store to read from a path iirc

[opal nimbus]

Hi everyone!
I'm working on a project that needs to parse sign block-entity data.
My goal is to display them on a website with HTML, so the text, colour and glowing-ness need to be extracted, so I can insert it into the HTML template.
I used to do all the NBT parsing myself, with manual handling of edge cases, like some plugins putting data where it shouldn't normally be, according to the vanilla spec. But I try to handle these things as best as I can anyway.
I also support many Minecraft versions, all the way back to 1.13
So I've been wondering whether using Adventure would be beneficial for me here.
Could it handle (at least a part) of this parsing for me?

[forest meteor]

yo

my gson serializer is serializing weird for 1.21.4

{"color":"yellow","click_event":{"action":"open_url","url":"https://store.servername.net"},"text":"[➥ Store]"}```

look at this

click_event instead of clickEvent

url instead of value

[torpid robin]

yeah you're serializing to a newer version's format than 1.21.4

[forest meteor]

ye why is that

I just use GsonComponentSerializer#gson

[sand drift]

Well, what are you doing?

[torpid robin]

probably using a version of adventure not made for 1.21.4

[forest meteor]

its latest adventure

[sand drift]

Okay, but what binding?

because the output is version dependent

[forest meteor]

.. wym binding

i just use #serialize(Component)

[sand drift]

Okay, but, you are likely using adventure on some form of platform

[forest meteor]

oh paper

[sand drift]

what version of paper?

[forest meteor]

1.21.4 idk

latest one

but i relocate adventure

[sand drift]

So, you're not using paper adventure, you're likely using adventure-platform or whatever

[forest meteor]

yes

but thats not the problem tho

[rare sage]

it might be

[forest meteor]

im not sending it to the player yet

[sand drift]

Well, adventure isn't magical, it doesn't know what version of MC you're serialising for

[forest meteor]

how do i make it figure it out then

[sand drift]

so it's just going to use some default, given your complaint, the default is likely for the latest version

[rare sage]

the serializer serializes based on the current server's capabilities, native adventure integration does not have this problem, platforms kinda need to detect that

[forest meteor]

okay how do i help it out tho

like how do i provide a version somehow

idk much about server capabilities

[forest meteor]

but what i do know is that this is incorrect for my use

[sand drift]

I'd guess the question is, why are you using the gson serialiser directly?

idk how the version detection stuff works in adventure-platform, so, I'm kinda out

[forest meteor]

gson-serializer is separate module from adventure-platform

[rare sage]

i see adventure-platform tries to set the serializer version so it should be fine if the server is the version you claim it is

https://github.com/KyoriPowered/adventure-platform/blob/7ece7ac519f0c2dd00ad25205c28f50237d71a78/platform-bukkit/src/main/java/net/kyori/adventure/platform/bukkit/BukkitComponentSerializer.java#L64-L68
but this is indeed a question for #adventure-platform-mod-help

actually, yes you are using the serializer directly

that will always be at latest version

if you want the version specific serializer you need to get it from that class
but the fact you are manually serializing a component is kinda sus

[forest meteor]

oh perfect

I can use BukkitComponentSerializer#gson

[slate current]

Well, you can use adventure-nbt to parse the data if you aren't already - beyond that it depends on what exactly your system is; are you getting the data while the server is running? if so you could probably use NMS to load the NBT with Mojang's logic and avoid needing your own handling.
As for what adventure specifically can do, beyond the NBT bit you could use it to parse the components, but I think that's about it, it's not really meant for handling block entity data or anything like that.

[opal nimbus]

Thanks for the reply!
I didn't know there was an nbt deserializer 👀 Somehow I wasn't able to find it before...

I can't use NMS, because this project will have to work on other server software than just Paper
And it even has to work without a server at all, running completely externally.

[rare sage]

i mean, you can totally use adventure-nbt to handle block entity data if you so desire, it's just an nbt parser after all, if it can read it, you can fiddle with it

[opal nimbus]

I see

I do already have an NBT Parsing library I'm very happy with, but it doesn't do more than parsing the NBT, like properly deserializing everything into text components and such

I'd have to deserialize everything myself, manually

Which seems like a rather daunting task

[rare sage]

adventure is no different

[opal nimbus]

😔

[rare sage]

it just reads the binary and gives you a BinaryTag

[opal nimbus]

I see

I was hoping I could use it to handle the text component stuff for me

Because sign text has been changed many times over the versions
Sometimes even including JSON for extra "fun"

[slate current]

Is there no way to use e.g. the gson serializer detached from a server? don't actually know how it works internally

Or alternatively, just run a server as your backend

[sand drift]

I mean, your issue is DFU if your issue is the upgrade stuff

well, maybe; getting the structure is one thing, being able to load it on the server is another, as they just deserialise it into a json element tree and then run it through their codec

[opal nimbus]

I have indeed been suggested this:

But that sounded rather cursed to me, so I was hoping there'd be a better way than this

(BlueNBT is the NBT library I am using)

[opal nimbus]

But I don't want to Fix Up any Data
This needs to be able to run on essentially archived world files from all the way back to 1.13

[sand drift]

I mean, when you have the NBT one, transforming to gson and parsing it is doable, mojang already supports that through DFU codec funkery

[opal nimbus]

Or do you mean fixing up the data at runtime, without modifying it on the actual files on disk?

[slate current]

^ that's what the DFU is

[opal nimbus]

Aah, I see

I must admit I don't have any experience with it

[sand drift]

I mean, it really depends on what you want to do, part of the reality is that the format is evolving

if you wanted to use the servers built in serialisation stuff, you would need to run the underlying data through DFU somewhere so that it can parse it

[opal nimbus]

Myeah

[slate current]

It's Mojang's system, very rough explanation but basically just - all data is saved with an data version int, and when loaded the data is ran though a list of fixes determined by comparing the data's version to the current version to make it match modern format

[opal nimbus]

But there won't always be a server when my code runs

So I cannot rely on it

[obsidian nimbus]

flyway but for json?

[opal nimbus]

It's essentially like a CLI tool that runs on the world files themselves directly

[sand drift]

Well, Mojang does have some gross codec stuff to convert NBT to Json 😄

But, I mean, once you have it in json iirc adventure is fairly adept at decoding it

[opal nimbus]

Ahh that makes sense
I've done similar stuff in the past: https://github.com/TechnicJelle/BlueMapSignExtractor/blob/ffb576e2e8d4ee9b468d5ab629e96fc1b252350c/src/main/java/com/technicjelle/bluemapsignextractor/common/ChunkClass.java#L29-L55

[sand drift]

There is DataConvertor, idk how detatched it is from MC

not entirely, because ofc it needs access to mojang stuff to deal with parsing data, NBT, etc

[opal nimbus]

Then I might try making some classes that mirror the various NBT layouts of the sign data over the versions
And then parse the nbt according to the correct version, like I'm doing here
And then I can convert that to json, and then feed it into adventure-lib...

[short kelp]

is it normal that mm crashes with negative rainbow phase? 🤔

<rainbow:-1>1234567890</rainbow>

    at net.kyori.adventure.util.HSVLikeImpl.requireInsideRange(HSVLikeImpl.java:61)
    at net.kyori.adventure.util.HSVLikeImpl.<init>(HSVLikeImpl.java:36)
    at net.kyori.adventure.util.HSVLike.hsvLike(HSVLike.java:51)
    at net.kyori.adventure.text.minimessage.tag.standard.RainbowTag.color(RainbowTag.java:105)
    at net.kyori.adventure.text.minimessage.tag.standard.AbstractColorChangingTag.apply(AbstractColorChangingTag.java:126)
    at net.kyori.adventure.text.minimessage.MiniMessageParser.handleModifying(MiniMessageParser.java:271)
    at net.kyori.adventure.text.minimessage.MiniMessageParser.handleModifying(MiniMessageParser.java:273)
    at net.kyori.adventure.text.minimessage.MiniMessageParser.treeToComponent(MiniMessageParser.java:248)```

🎯

[quaint monolith]

As mentioned in #paper-dev you could argue that MM itself should be throwing some kind of parsing exception instead but it should be expected for it to fail when you feed it invalid data

[robust wharf]

Normally MM just ignores the tag for exceptions iirc

Idk, open an issue I guess and we can see

[quaint monolith]

Hmm, I dunno if it would be better to just strip the rainbow tag in that situation

More robust but more confusing

[devout sage]

From a design standpoint, if I'm hotswapping between two boss bars (ie removing one from a player and adding another one in the same tick), is it better to create brand new instances or modify the existing boss bar with new color/text/progress

[cloud vapor]

Don't think it really matters too much, whatever you prefer really

[cloud vapor]

Yes we shouldn't be erroring here - please open an issue!

[slate current]

👀, as far as I can see data is auto wrapped when saving, but isn't unwrapped when loading - but maybe there's a feature for that somewhere that I'm missing?

[kind jungle]

How does logging components to a console work under the hood in adventure?

how is Windows support and checking terminal capabilities done? How do you convert minimssage into something a console can understand? What does the fallback process look like? (Full color range -> 256 colors -> ANSI colors -> strip colors?)

[cloud vapor]

We have an ANSI serializer that does all of that

[robust wharf]

fyi, motd preview on the minimessage webui is currently broken

[robust wharf]

fixed

[worthy adder]

hey, can i use <click:suggest_command ...> and replace placeholders inside the command thereafter?
So <click:suggest_command:'/snapshots restore <username> <snapshot_unique_id>'>Restore</click>, and replace the <username> etc to the actual variables?

[cloud vapor]

yes

use preparsed placehodler

[worthy adder]

do u have an example or smth?

[keen wraith]

I'm here for an issue I'm encountering:
you see, i have to modify my shadow color in order for my item description to look clean, so I use <!shadow>. but I want to reactivate the shadow, so I just put a <reset> tag, but the shadow doesn't go away like it's supposed to. am I doing something wrong here, or am I just dum ?

[wet kestrel]

<username> could be done as Placeholder.parsed("username", "mbaxter") for example

[sweet hornet]

You probably don't want to use <reset>. Please provide a full example of MM string with your expected result

[keen wraith]

<shadow:#FF08FFFE><#FF08FF>Mining Fortune: <lang:space.71><reset><#53F52F>+65<lang:space.7>

[sweet hornet]

Placeholder#unparsed sounds more appropriate, since it is for a command

Oh wait no, I just inverted in my head

[worthy adder]

worked, thank you!

[sweet hornet]

Tbh, I find those two names (parsed vs unparsed) a bit confusing when looking at them from a distance. Because on one hand, parsed inserts the literal string into the parser, but on the other hand, if you want to get the literal value untouched, you want unparsed. Both have something to do with raw/literal values, but not in the same way

Maybe it is clearer than I think for native English speakers tho, and it is only amplified by the language barrier

[wet kestrel]

You want the string parsed by the parser? Parsed

You want it not parsed by the parser, and untouched? Unparsed

[keen wraith]

I fixed it, my code was just not adding the tag :/

[sweet hornet]

Yes but look at the above example : We want the player name not to be "parsed", we want it to stay as a string (so the first thing to come in my mind is unparsed). That's why

[wet kestrel]

You do want the player name parsed, as a part of the click event.

[sweet hornet]

I think the misleading part is the fact we don't want to parse tags inside the string (in that case, there should be none, but in a general case)

But I agree about the correctness of the names from the parser point of view

And maybe it's just me who is having trouble every time I need to choose, and I always go inside the code of either one to identify which one I need

[fiery oracle]

(i also find them confusing, as a native speaker)

it's probably an issue from parsed/unparsed being able to be both "should be parsed/unparsed" and "is parsed/unparsed"

[rare sage]

The fact that this discussion is currently taking place is indicative in and of itself that the names are confusing

[sterile rampart]

the split between the two types of tags is imo a design flaw anyways

[jolly sage]

I would like to put some translatable messages into .properties or .json file (like minecraft does in the lang folder) but am confused by the docs and all methods that exist, for example there is this one but I'm unsure how to get a StringBased translation store and also there is no other method that takes a resource bundle

I took this code from docs.advntr.dev Localization:

Is using resource bundles the right way to do this anyway? i did this by right clicking on resources then new then Resource Bundle in Intellij

[cloud vapor]

A component translation store isn't a string based translation store

You want either TranslationStore.messageFormat or MiniMessageTranslationStore

[jolly sage]

I thought Component.translatable method works only with a component translation store :/
Still, how do I put in a resource bundle? Or any other way of reading keys and appropriate translations from various files

The method is still not showing 😦

[cloud vapor]

Again, you need a TranslationStore.StringBased - you're just not using that in your type

[jolly sage]

OH that's it

i'll try now

[jolly sage]

how do I refer to these keys now though?
I tried everything

[cloud vapor]

did you register your translation store

[slate current]

Should I move this to #adventure-contrib? I just want to know whether I'm good to PR this, and if so, whether there's any specific way you'd want the option implemented (I assume not by default?) - currently if you save a heterogeneous list to file and then load it, you get basically a different list back, and handling unwrapping it manually is not great

[jolly sage]

I'd say so, and yeah i called this method in onEnable (is this correct?)
in the .properties file Intellij is showing it as "unused property"

[jolly sage]

omg i finally got it, it just needed some-key and not these namespace keys whatever, saw it here

thanks for the previous help 🙌

it's showing me the default one instead of localized :( nvm i just got confused by Locale.FRANCE and Locale.FRENCH (tested by setting french ingame)

[vivid bobcat]

Hello, I have my own MessageTranslator class that translate messages into components, everything is OK. All the messages are mapped on a Map<Locale, Map<String, String>> messages; that I load from a json:

{
  "messages": {
    "en_us": {
      "hello.player": "<green>Hello, <player>!"
    }
    "pt_br": {
      "hello.player": "<green>Olá, <player>!"
    }
  }
}

But I'm wanting to do something spicy, I want to create a TagResolver "ref" to reference messeges between themselves, example:

{
  "messages": {
    "en_us": {
      "player": "player",
      "command.ban.syntax": "/ban \<<ref:player>>"
    }
  }
}

How can I create a TagResolver like that? should I implement Modifying interface? I wanna make it even more spice, if possible, by adding some arguments, for example: ref:player:to_upper_case or some regex replacements.

[alpine marlin]

Does someone know why the date does not get displayed in the message?

[last horizon]

.

@alpine marlin

[alpine marlin]

that's kinda wierd, but thank you 😄

[last horizon]

it was explained a few messages after mine if you are interested about the why

[fiery oracle]

Is there any particular reason you're not using the global translator?

[rare sage]

https://docs.oracle.com/en/java/javase/24/docs/api/java.base/java/time/format/DateTimeFormatter.html#getZone()

so you need to pass a zoned temporal accessor, Instant is not it

[vivid bobcat]

global translator? it returns a MessageFormat, seems that it isn't good for me, I'll have to make the messages args {0} {1}, unlike MiniMessage

[rare sage]

what

you can totally register a minimessage translator into the global translator

[fiery oracle]

you can use whatever format you want with the global translator, with the translation store stuff

https://jd.advntr.dev/api/4.23.0/net/kyori/adventure/translation/TranslationStore.html

for minimessage specifically:
https://jd.advntr.dev/text-minimessage/4.23.0/net/kyori/adventure/text/minimessage/translation/MiniMessageTranslator.html
https://jd.advntr.dev/text-minimessage/4.23.0/net/kyori/adventure/text/minimessage/translation/MiniMessageTranslationStore.html

[vivid bobcat]

can I use that for lores? Because currently my MessageTranslator class have the method .translate(TranslationRequest) that returns a TranslationResult, then I can do toComponentList() or toSingleComponent() or toLegacy() toPlainText() etc

[quaint monolith]

ItemStacks can't have server-side translations

[fiery oracle]

the automatic translation of the global translator does not operate on items (at least on paper, not sure about other platforms)
but you can still manually invoke the global translator

[quaint monolith]

Well, yeah, I suppose you could manually replace the component with the translated version on the server

[fiery oracle]

as for the list thing, not exactly sure what you mean by that
you pass a component to the global translator, and it renders it using its registered translators

you get a component back, with any translatable components translated if a registered translation is found

you can still have your overall system, but it'd likely be easier to let the global translator handle the key + locale -> translation part

(and the global translator will auto translate messages sent to players, along with some other areas, just not items)

[vivid bobcat]

I didn't get It, can you give an example?

[fiery oracle]

wdym

https://docs.advntr.dev/localization.html

[vivid bobcat]

how to use it? it would be something like that?
player.sendMessage(Component.translatable("hello.player"))?

[fiery oracle]

ye, that's all you'd have to do, if your platform supports the global translator

if you have arguments you'd have to pass them as well

[vivid bobcat]

currently on my system all the args are on a AgnosticMessageContext, it has a List<TagResolver> field

[fiery oracle]

you can pass tagresolver arguments with the minimessage translator

[fiery oracle]

the javadoc of the 1st link in this message explains some & shows an example of passing arguments

[ripe tulip]

is it possible to get the <namespace> placeholder to be filled in this minimessage string, or should i rework how i am making this message?
<hover:show_text:"<ash>Click to view info."><click:run_command:"/module <namespace> info"><tyrian><module>

return Component.translatable(
        "message.module.module_name",
        Argument.string("namespace", namespaced()),
        Argument.string("module", formatted())
);

<module> works but i assume <namespace> isn't because its nested in a click tag

[wet kestrel]

I don't know what Argument.string is doing but it should be returning a tag resolver that produces a pre process parsed tag if you want <namespace> to get processed, like how Placeholder.parsed works.

[ripe tulip]

i messed around with arguments until i got it to work, so could i ask if this looks fine to you? is this what you meant

return Component.translatable(
        "message.module.module_name",
        Argument.tag("namespace", Tag.preProcessParsed(namespaced())),
        Argument.string("module", formatted())
);

[quaint cedar]

hi, i'm having issues using mimimessage to create messages, font and hover and click tag do not work at all

Context:
Version:
This server is running Paper version 1.21.8-6-main@782ce95 (2025-07-18T20:26:58Z) (Implementing API version 1.21.8-R0.1-SNAPSHOT)

Code:

       Component component = MiniMessage.builder().build().deserialize(builder.toString());
        if (player != null) {
            getAudiences(plugin).player(player).sendMessage(component);```

Text sended:
<font:minecraft:alt> Font</font>
<hover:show_text:'<gray>This is a tooltip!</gray>'> <gold>Hover over me</gold> </hover>

plugin.yml libraries

libraries:

• net.kyori:adventure-platform-bukkit:4.4.0
• net.kyori:adventure-text-minimessage:4.23.0```

pom.xml

        <!-- MiniMessage -->
        <dependency>
            <groupId>net.kyori</groupId>
            <artifactId>adventure-text-minimessage</artifactId>
            <version>4.23.0</version>
        </dependency>
        <dependency>
            <groupId>net.kyori</groupId>
            <artifactId>adventure-platform-bukkit</artifactId>
            <version>4.4.0</version>
        </dependency>```

Am I missing something? Any tips would be really appreciated 🙂

[wet kestrel]

Paper does not require adventure-platform. Just use player.sendMessage and you won't have this issue.

Alternatively, if you insist on using the community-maintained -platform, you can find code that seems to work in a PR on the repo. The current release does not support recent changes (1.21....6? 5? I forget... and up)

[quaint cedar]

Yea, I'm supporting spigot too.. maybe I can walkaround and use the Gson serializer and directly send the message as a packet

Or using the Bungee serializer to turn into bungee components

I'll try to find a solution, thanks

[wet kestrel]

Probably easiest to just use the PR'd code, honestly.

But, it's also a great time to start migrating from a server that has a hand-countable percentage of servers. 🙂

[cloud vapor]

Yes that looks correct! A pre process parsed tag inserts the contents of the placeholder into the literal string before it goes into fun component land, which means the click event works (bc click event eats normal strings, not components)

if that makes sense

[worthy adder]

is it possible to add a translation to my holograms plugin so that i can register a minimessage key via plugin, e.g. <t:TEXT>, and then the player will see the translation of TEXT depending on their language? So without packets

[jolly sage]

are you looking for https://docs.advntr.dev/minimessage/format.html#translatable ?

[worthy adder]

kinda, but i need to use my own translation system for this, not the default minimessage translator with a file or smth

[cloud vapor]

Translatable components are what you want, specifically coupled with this: https://docs.advntr.dev/localization.html#using-adventure-s-localization

[grim bear]

I use CarbonChat which has a feature where you can link items in chat as a hover event. Is there a way I can find these hover events in a message, and edit them to also show the quantity of the item that was linked? Or is that data lost at this point

[sterile rampart]

That feels possible, might be worth talking to Josh directly though and seeing if you can just modify carbon to do that when it's inserting the hover events in the first place

[grim bear]

I raised a suggestion for it a while ago but it hasn't been looked at, so I'm thinking about just adding it in my plugin in the meantime

the suggestion was for an ItemLinkEvent which would have info on the item, to be more specific: https://github.com/Hexaoxide/Carbon/issues/658

[obsidian nimbus]

you may just fork carbon at that point

[grim bear]

probably harder than what I want to do here, assuming it's possible

[fiery oracle]

https://jd.advntr.dev/api/4.23.0/net/kyori/adventure/text/event/HoverEvent.ShowItem.html#count()

[opal nimbus]

this is indeed the path i have taken, and it seems to be working very nicely!
currently, i'm just serializing the text with the plain text serializer, and my signs are looking very pretty
except they don't have any colour...
so now i want to properly serialize the net.kyori.adventure.text.Component into an HTML string, instead of using the plain text serializer
does anyone have any hints on how to beging doing that?

for starters, i can't seem to find a getter to access the actual text of Components

[sand drift]

Not all components are text components

[opal nimbus]

oh

[sand drift]

You would generally want to take a look at the component serialisers and how they work

[opal nimbus]

i have been looking there a bit

but it's a loot of code

[sand drift]

Well, yea, it's a lot of fun

[opal nimbus]

i thought i'd try being inspired by the ansi encoder

[sand drift]

You'd mostly want to be walking over the component and their children

[opal nimbus]

but eehh

mhm

yea

but how do i get the actual content

i couldn't discern that

[sand drift]

You would basically need to branch depending on the type of component

[opal nimbus]

in the plaintext serializer i see a StringBuilder

[sand drift]

Well, yea, as that's mostly just using a flattener

[opal nimbus]

but i can't actually see where it's calling that append

[sand drift]

that's the flatteners doing

[opal nimbus]

even in here, i don't see it

or does it happen in the .handle( method?

[sand drift]

https://github.com/KyoriPowered/adventure/blob/2f48b6b0cd9332781f7ec724012391f4a70c663b/api/src/main/java/net/kyori/adventure/text/flattener/FlattenerListener.java#L35

[opal nimbus]

is that where the sb::append function gets actually called?

[sand drift]

I would imagine that flattener.handle calls the component method there

[opal nimbus]

but intellij can't find any implementers of this interface

so i can't track it down any further

[sand drift]

I mean, an implementation of that is the sb::append call

[opal nimbus]

whaoh

okay hold on i need to process that

functions, like sb::append, can implement an interface?

[sand drift]

Yes

[opal nimbus]

wow

i did not know that

[sand drift]

it's a functional interface, i.e. an interface with only one method, which can be implemented using a lambda

[opal nimbus]

😵‍💫

[sand drift]

https://github.com/KyoriPowered/adventure/blob/main/4/api/src/main/java/net/kyori/adventure/text/flattener/ComponentFlattenerImpl.java

has the "basic" flattener

[opal nimbus]

ooh that's where i have to look

[sand drift]

I mean, I'm not ultra familiar with this stuff, just cursory glancing

if I was going to start on something that poops out html I would basically start with a span tag, and every time it calls pop/push style, I'd basically de/nest a span tag or something with that info

[opal nimbus]

yeah that's also what i was planning

[sand drift]

would probably be horrible but I would hope that would give me enough familiarity of how it works to be dangerous

[opal nimbus]

but didn't know where to start at all

thank you for the information!

i really appreciate it 💜

[rare sage]

i guess you can look at how the minimessage web viewer does it

[opal nimbus]

for tonight, i'm going to just use this:

public class HTMLComponentSerializer {
    //TODO: Do not use this Plain Text Serializer.
    //This needs to properly support nesting.
    private final static PlainTextComponentSerializer PLAIN_TEXT_COMPONENT_SERIALIZER = PlainTextComponentSerializer.plainText();

    private static final HTMLComponentSerializer INSTANCE = new HTMLComponentSerializer();

    public static HTMLComponentSerializer html() {
        return INSTANCE;
    }

    public String serialize(Component component) {
        //Get Text
        final String text = PLAIN_TEXT_COMPONENT_SERIALIZER.serialize(component);

        //Get Colour
        TextColor componentColour = component.color();
        final String colour;
        if (componentColour != null) {
            colour = componentColour.asHexString();
        } else {
            colour = null;
        }

        //Generate HTML
        if (colour != null) {
            //The colour for this specific line has been overridden
            return "<span style='color:" + colour + ";'>" + text + "</span>";
        } else {
            //There are no overrides, so glowing here is not necessary. It is handled by the parent sign
            return "<span>" + text + "</span>";
        }
    }
}

i will improve this to actually handle nesting later...

it's already starting to look really good, though!

using adventure was a very good decision

this library is awesome!

i was able to remove a lot of custom parsing code \o/

in exchange for custom encoding code

but i had to do that anyway

even for my previous version which didn't use adventure

[opal nimbus]

i have now expanded it to this: https://github.com/TechnicJelle/BlueMapSignExtractor/blob/4645b538257a81a5ac6d28746868155ea3f69a60/src/main/java/com/technicjelle/BlueMapSignExtractor/HTMLComponentSerializer.java
what do y'all think?

as far as i've seen in my tests, sign text is generally just normal TextComponents, so the instanceof TextComponent that i'm doing isn't the most evil think, i think...

although it may also have Translated Components

i think i have a region file somewhere that has those, so i'll throw that into my testing suite now and see how it deals with it

[wicked torrent]

you probably want to sanitize that output

the serializer in squaremap api uses some google library

[naive sphinx]

Is there adventure to html converter?

[devout sage]

Probably not what you want, but the minimessage format is xml

If you're asking about actually replicating minecraft ui formatting on a web browser, uh, hope you can code

[naive sphinx]

I have text with color: and etc

Which I use in website and server

In website I currently stripr all the <>

Maybe it's possible to convert those tags to html elements with colors

[cloud vapor]

Scroll like two messages up lol

and also yes, adventure-webui serializes to HTML via a public API you can use or you can see how it does the serializing yourself

[naive sphinx]

Yea, i'm looking more for a php converter, for websites

[cloud vapor]

PHP just spits out HTML

[robust wharf]

Basically, we don't wanna reimplement the MiniMessage parser in other languages in fear of stuff not being aligned, so you should just use the API to get the html

[hoary plume]

I'm really confused regarding NBT in Adventure.

I got the exception that I need to use a Compound Tag for the "additions" in a dialog, but I can't just use CompoundBinaryTag.builder()...build() because it wants me to provide a BinaryTagHolder.

Looking at the Javadocs, it tells me to check net.kyori.adventure.nbt.impl for a platform agnostic implementation, but this isn't linking to anything and also doesn't make it very clear where exactly I should check (I assume one of the GitHub Repositories, but the main adventure one doesn't seem to have such a path anywhere? Is it referring to the archived NBT repo?)

tl;dr: How would I convert a CompoundBinaryTag into a BinaryTagHolder?

[cloud vapor]

it's referring to the adventure-nbt module

[hoary plume]

That isn't helping me much here

[cloud vapor]

net.kyori.adventure:adventure-nbt

https://jd.advntr.dev/nbt/4.23.0/net/kyori/adventure/nbt/package-summary.html

[hoary plume]

Well....

[cloud vapor]

it's an nbt library

[hoary plume]

Does it have the source anywhere to check?

[sweet hornet]

https://github.com/KyoriPowered/adventure/tree/main/4/nbt/src/main/java/net/kyori/adventure/nbt

[cloud vapor]

source for what?

[hoary plume]

The stuff it mentions in the docs but doesn't bother to link properly

[cloud vapor]

Feel free to submit a PR to correct the docs

[hoary plume]

And this btw doesn't have the impl package either

[cloud vapor]

I've linked you to the javadocs for adventure-nbt and you've also been linked the source too

do you have a specific question?

[sweet hornet]

I think that jd comment is just wrong, and the impl subpackage does not exist. However, ...Impl classes are present in the nbt package

[hoary plume]

Exactly that

The docs imply there is a impl package in the NBT library when there isn't

[sweet hornet]

https://github.com/KyoriPowered/adventure/commit/cb2cb2b3cb64b2204f00553585eff603e5d537a2

[cloud vapor]

Yes and you're free to submit a PR that fixes that if you would like, otherwise you're just complaining when we're waiting here to help you with any actual problems you have :p

[sweet hornet]

Since this commit, the impl subpackage was merged into its parent

[hoary plume]

So.
I apparently need to provide a Compound tag according to the exception I've received.
Tho, in the method where I need to do that does it require me to provide a BinaryTagHolder, which seems to be some form of common interface here?
My current question is, what the best aproach would be to make a Tag that satisfies the BinaryTagHolder requirement while also being a compound tag.
Would I need to make my own implementation (Like the CompoundTagImpl example), or would there be a simpler aproach available in Adventure.
I use adventure through paper, if that is of any significant importance.

[sand drift]

BinaryTagHolder is something that just holds a tag, it's not a tag

it looks like there is a method on BinaryTagHolder that lets you create an arbitary holder

[cloud vapor]

You can create an instance using BinaryTagHolder#binaryTagHolder(String) where the parameter is the snbt

[hoary plume]

And I assume I would call some asString method or similar in the compound tag to convert it into a valid snbt string?

[cloud vapor]

See BinaryTagIO

[hoary plume]

This is so confusing....

I don't get how BinaryTagIO is helping me here....
Like if I look at Readers I only see methods for getting CompoundBinaryTags, which is something I can get using the Builder already...
And when I check the Writer I can see void methods that need Paths or OutputStreams or similar, which I doubt is what I need here?

[sage river]

I suppose there could be some kind of DirectBinaryTagHolder that would hold BinaryTag instance in adventure-nbt library to make it simplier 🤷

[cloud vapor]

You can write a created binary tag to whatever output stream way you want to get a string

plenty of different methods to turn output streams to strings and w/e

[hoary plume]

Seems kinda overcomplicated to me...

[cloud vapor]

Not really, it's kinda Java IO 101

But like this is just a suggestion of what library to use, obvs you can use whatever nbt library you want as long as it can output snbt you can throw into BinaryTagHolder

or make your own one, or use nms, or whatever really

[hoary plume]

All I seriously want is, to have one single integer added to a BinaryTagHolder, so that I can later extract it for further use

Nothing more

[cloud vapor]

why?

can't you just use click callbacks and just hold the reference to whatever int you want during the dialog?

[hoary plume]

I'm using Paper's custom click event

[sand drift]

I mean, from a 2 second glance, you'd create the compound tag, then use TagStringIO and toss the output into a BinaryTagHolder?

[cloud vapor]

why?

[sage river]

BinaryTagHolder.binaryTagHolder(TagStringIO.tagStringIO().asString(yourCompoundBinaryTag))? doesn't seem that much overcomplicated, though DirectBinaryTagHolder would be nice

[cloud vapor]

just use ClickEvent.callback

[hoary plume]

Seems like there is only a get() and builder() method, but I get what you mean

[sage river]

where

[hoary plume]

I think the event is a bit more clean than having the entire handling be done in a callback. Even if I throw everything in a separate method.

[hoary plume]

TagStringIO

[cloud vapor]

so clean you have to include an entire nbt library haha

[sage river]

you have outdated adventure-nbt

[hoary plume]

Blame Paper on that

[cloud vapor]

don't reinvent the wheel, ClickEvent.callback is a user-friendly alterantive to doing it yourself

[sage river]

what

[cloud vapor]

Paper doesn't ship adventure-nbt

[sage river]

Paper doesnt include adventure-nbt

[hoary plume]

K...

[sage river]

and it doesnt have to match adventure api version

[hoary plume]

Then screw this honestly... I won't shade in/download nbt on runtime just for this...

Wanted to use the Event but guess this stuff is just too annoying to setup properly.

[cloud vapor]

Good idea, you're saving yourself a world of hassle by using the nice clean click callback system we made haha

[hoary plume]

Not too much

But whatever

[cloud vapor]

click callbacks are a wrapper around the custom click event - so it does the nbt, serialization/deserialization, filtering, event handling all for you

if you wanna do that all yourself then go ahead we're not stopping you

even has more logic for custom expiry and uses as well

[sweet hornet]

Just a side question on this topic, I am just wondering, when showing a dialog to a user, with multiple click actions (with callbacks), he only click one of them, and we show him an updated version of the dialog. Aren't we "accumulating" callbacks into memory, that will never be used?

I.e., is setting a really short-lived expiration a good practice to prevent this?

(a few minutes)

(and of course single use, but that only work for the one that is clicked)

[cloud vapor]

hm - i mean i cant see it actually being a problem

but yeah maybe - or we could have a "group" system so that other callbacks with the same group expire when one does or smth

[sweet hornet]

I did not dig into the implementation, but I guess each time we send a callback, Paper needs to remember it somewhere, in a Map or something, no?

[cloud vapor]

yeah

i cant remember how paper implements it either

[sweet hornet]

So, in a scenario of a user-controlled trigger (like, a dialog showing/updating based only on user input), wouldn't it be possible for a malicious user to "fill" the memory with unused callbacks?

[cloud vapor]

you'd probably want a sensible expiry anyway

[sweet hornet]

(of course not limited to dialogs, it would also apply to chat-based callbacks)

Yes, but the default one being 12 hours, I something stumble across code where I forgot to lower it, and I always had this question of whether it would be a possible attack vector or not

I guess it is probably not the most obvious way an attacker could think of, I admit it

[cloud vapor]

yeah idk, hard to say really

i do remember being not very impressed with paper's implementation of the click callback stuff, but then my ideal solution is just throw it in a nice easy caffiene cache haha

[grim bear]

is it safe/recommended to store an object with Key fields in json files? or is it better to have them as strings and create keys from them when needed in runtime? it seems to work fine, but that doesn't always mean it's the best approach

the keys I'm saving are DamageTypes and EntityTypes

I'm using the following Gson serialiser in case that's relevant, because my object also contains Component fields

[cloud vapor]

I don't see any particular reason why not to

[grim bear]

as in, why not to do it the way I'm already doing it?

[cloud vapor]

Yeah

nothing wrong with using keys anywhere

[tranquil moth]

can legacy symbols be converted to minimessage format?

like the string
&cHello world
becomes
<red>Hello world
or will I have to do that manually?

[sand drift]

MM doesn't "convert" anything

you would need to deserialise it with the legacy serialiser and then throw it into minimessage

[wet kestrel]

Your earlier question was:

Is it possible to deserialize both legacy and minimessage format in the same string?

This is not going to work out happily. Like, there's no solution that results in predictability or joy.

If you have legacy stuff from an earlier version of the plugin, first do a one-time migration from legacy to minimessage. Then you can start using minimessage going forward. Don't mix and match.

[tranquil moth]

kk

is there a website or something I can use to copy n paste my messages to convert the legacy to minimessage? or do i have to manually find and replace?

(for that one-time migration)

[wet kestrel]

You should read both messages cat wrote instead of just the first one. The legacy serializer can deserialize to component, and then you can serialize to minimessage to save it.

[tranquil moth]

nvm its just like 4 things i have to change. thanks lol

Slightly confused now. For some reason my messaging system doesn't deserialize the minimessage string.

commandSender.sendMessage(MiniMessage.miniMessage().deserialize(message));

instead I just get the raw string.
(e.g. <red>Hellow world)

[sand drift]

Did you run the MM string through the legacy serialiser or something?

[tranquil moth]

nope

[sand drift]

probably want to print out the raw string before you deserialise it, hard to say with the info provided

[wet kestrel]

Also note that, if just doing default minimessage deserialization, you can do commandSender.sendRichMessage(message) which is waaaaay shorter.

[tranquil moth]

nvm im an idiot. i forgot to rebuild the plugin-

[tranquil moth]

oh thanks

[wet kestrel]

And even sendRichMessage(message, TagResolver) if adding any, actually! Forgot about that one. :3

[tranquil moth]

what do tags do?

[wet kestrel]

So basically, unless you're doing a non-default MM instance (which is rare) you can save a lot of time.

[tranquil moth]

oh you mean the <xyz> tags

[wet kestrel]

Yes

[tranquil moth]

ah ok that makes sense

[wet kestrel]

commandSender.sendRichMessage("Hello <name>!", Placeholder.unparsed("name", "SpyCoderX")); would send you "Hello SpyCoderX!"

[tranquil moth]

oh cool

[wet kestrel]

Check the MM docs/javadocs for info on that.

[safe epoch]

hi, i tried using MiniMessageImpl#deserialize in every tick for updating scoreboard for each player and it just generate too much memory usage. i need to be able to format text and minimessage does the job
is there any reason it should use so much RAM? any alternavite or suggestion please?

[grim bear]

is there a way to translate DamageType? I can't seem to call Component#translatable on them

[fiery oracle]

damage types don't have associated translation keys

what exactly are you trying to do?

[grim bear]

Display top 5 causes of death in a hologram. I have a damage type key stored for every death

I guess I can just map my own translations

[cloud vapor]

paste the full spark log

[final nimbus]

hello guys. I'm right now trying to achieve text with translations. I've got a .yaml file where I get my translations on MiniMOTD format.

game:
  round:
    started: "<white><gold><b>The game has started!</b></gold></white>"

  over:
    no_winners: "<white><red><b>Game over!</b></red> No winners this time.</white>"
    single_winner: "<white><green><b>Game over!</b></green> <aqua><b>{player}</b></aqua> is the winner!</white>"
    multiple_winners: "<white><green><b>Game over!</b></green> Multiple winners survived!</white>"

Doing

player.sendMessage(
        Component.translatable("game.scoreboard.title"))
  );

works neatly and as intended. but I'm trying to add tags. for example in game.over.single_winner, how could I dinamically change {player} to whatever I want via code?

I've seen I could tweak the deserializer to achieve my result, but I'm unsure how I could go on about it

any help would be greatly appreciated!

[fiery oracle]

by "MiniMOTD format", do you mean minimessage?

minimessage has its own system for dynamic replacements:
https://docs.advntr.dev/minimessage/dynamic-replacements.html

[final nimbus]

yup! sorry, matter of habit

[final nimbus]

Interesting, would there be any way to do it from the Component?

[fiery oracle]

are you registering your translations using the minimessage translator stuff?

[final nimbus]

sorry lol

[final nimbus]

I'm using a custom translator

[fiery oracle]

yes, but is it using the minimessage translator stuff?

https://docs.advntr.dev/minimessage/translator.html#creating-a-minimessage-translator
like, are you extending MiniMessageTranslator?

[final nimbus]

oh s- I see now what you mean

[fiery oracle]

or using a MiniMessageTranslationStore?

if either of those 2, what you sent is how you'd specify custom tag resolvers, yes

[final nimbus]

yeah that's the problem I was only implementing the interface

yeah dude you're 100% right, I feel so stupid right now lol, thanks!

[fiery oracle]

there's some explanation + an example on this javadoc page
https://jd.advntr.dev/text-minimessage/4.23.0/net/kyori/adventure/text/minimessage/translation/MiniMessageTranslator.html

[final nimbus]

okay that's just what I needed

I can't believe I was stuck on hours for this lmao

thanks man

[white sundial]

Does <newline> supposed to work in item lore?

Something seems off

Oh, discord search tells me it isn't supported in lore. Suggestion: make the web editor not imply that it does work.

[cloud vapor]

Yeah it's tricky, there's no nice way of actually making the webui imply it doesn't work unfortunately

[white sundial]

I guess a small red text on the top could do fine when the input string contains the tag

Or in the docs, just put it somewhere, I think it is an important note

[cloud vapor]

Feel free to submit a PR!

[tranquil moth]

Or you could create a function which splits a component at every newline, returning a list of components?

[rare sage]

Easier said than done, but feel free to submit a PR for that too!

[cloud vapor]

it's much easier to just have a list in your config/whatever

[tranquil moth]

Wdym? Wouldn’t you just serialize each component, split them at any newline
Then just get the color at the end of the previous line to copy to the following one, and reserialize it into a list of components?

[sterile rampart]

serialize to what, json?

[tranquil moth]

String

Minimessage

[rare sage]

no, the process is more complicated than that due to the tree nature of components

[sterile rampart]

yeah you have to maintain the full tree, no real benefit to serializing when all the info exists in the deserialized tree anyways

[tranquil moth]

Well, if you are doing it as a component tree, you just need to split components which have a newline in them into two components with the same position in the tree and same properties (color, style, etc.)

[rare sage]

yes, which again is easier said than done

no-one has PRd a fully working solution to adventure and many have tried, otherwise it'd be in the api

[tranquil moth]

Hmm

[rare sage]

text("green text", GREEN)
  .append(text(" red text in the first line\nred text in the second line", RED))
  .append(text(" also green in the second line"))

is enough to give headaches, and it's a single level deep

[sand drift]

You can't just cleanly split, you need to iterate over the tree and create new trees with the splits that you need

[rare sage]

let alone when a single component in itself is a single newline that could have its own style for whatever reason

it's not an impossible problem, mojang has solved this issue on the client text decomposition :p but it is not a trivial enough problem for the contributors' free time

[tranquil moth]

I see

From what I can tell, you’d only need to iterate over the tree and replace single components with newlines in them into multiple components with the same properties, but only the last one keeps the “extra” component list.

[cloud vapor]

and all of this is why i always say "just put a list in your config" haha

[tranquil moth]

lol. I love making a more advanced solution to a simple problem XD

[rare sage]

not necessarily only the last one would have the children

if you have multiple styles spanning multiple lines and depths

[cloud vapor]

i dont even think it's a nicer solution, i would hate being an end user and having to write everything on one long line bc you made a fancy splitting solution

[tranquil moth]

No because the whole point of newline splitting would be so you can do this:

Component.split(“””
This
can
split
“””);

and that would allow for writing things like lore in one component .

[rare sage]

sure but if you are taking it from a configuration file, unless you allow for both options you are also forcing the user to not have them tidy if they desire ¯_(ツ)_/¯

[tranquil moth]

🤷 I see problem, I solve problem. The user can do what they want lol

[rare sage]

they can't do what they want if you don't give them the freedon

[sterile rampart]

i mean config file is easy, it's just like:

my_item:
  lore: |
    <red>first line
    <bold>second line

[tranquil moth]

But… wouldn’t it be a function the user calls in their code?

[rare sage]

by user i mean end user of the application using adventure, where the application takes the "lines" from a configuration file the end user configures

[tranquil moth]

Ah

[rare sage]

which is why kezz keeps saying "take a list in your configuration file"

[sweet hornet]

I could just see this being slightly surprising for a user who might expect the second line to be red as well, because there's no </red> on the 1st one

[rare sage]

which is the simple enough solution

[sterile rampart]

it should be

[sweet hornet]

Huh

[sterile rampart]

read lore as a single component, then split it

[sweet hornet]

Wait, what was the point of your example?

[sterile rampart]

the split should preserve style across lines

[sweet hornet]

Oh ok

Yes yes yes

I thought you were illustrating the way we could ask the user multiple lines of lore, then parse each one individually

But for your example to work, we would need the holy graal Component#split

[tranquil moth]

Imma think on this

[cloud vapor]

and the alternative that doesn't require a WHOLE bunch of code lol

my_item:
  lore:
    - <red>first line
    - <red><bold>second line

[tranquil moth]

When you encounter a newline, wouldn’t you just copy the current list of colors and styles in order, exit out of the current json component, and create a new stack of components with the same styles and order, but with their text set to “”?

And then youd just have to integrate the remaining unsplit/unprocessed components into the newly created one.

[sterile rampart]

"just" is doing a lot of work there

[rare sage]

this is why i keep saying it's easier said than done, lol

the theoretical process is easy to think about at a high level

but actually putting in the time and writing all the gross state-tracking code is the challenge

[tranquil moth]

Couldn’t you copy the current node’s parent tree but skip anything before it and then set all of its parents text to “”?

[rare sage]

components don't keep track of their parent

do you know what "easier said than done" actually means btw?

[sterile rampart]

if, instead of spending all this time talking, you went and tried to implement something, you'd understand the issues way better

[tranquil moth]

I would but I’m at the pool rn

[wet kestrel]

Go enjoy the pool then

[tranquil moth]

XD

[white sundial]

I only wanted newlines to work in lore because I need to create some idiot/crowdin format proof config. So every lore would be just one line and splitting on new lines. But yeah this seems kinda complicated to do. Instead I decided to use wrapping which also seems kinda complicated to do, but found a gist which looks like is working perfectly fine or at least I couldn't break it yet
https://gist.github.com/Minikloon/e6a7679d171b90dc4e0731db46d77c84

[white sundial]

Will do once I get to it, hopefully for both places

[final nimbus]

hello everyone! I'm using currently Translation components and I'm trying to use them for Item's names. Here's a minimal working example of my code so far

ItemStack sword = new ItemStack(Material.DIAMOND_SWORD);
ItemMeta meta = sword.getItemMeta();
meta.itemName(
  Component.translatable("event.give_special_item_sword")
);
sword.setItemMeta(meta);```
However, in game, they only show the raw string `event.give_special_item_sword`. what do I have to do in this case?

I'm using a custom translator btw, and it seems to be working well so far

[sand drift]

Item components can’t be translated as that would require modifying the item and would break a bunch of stuff

[final nimbus]

hm, I suspected something along the lines of that, thank you!

[grave smelt]

How can I create a system where I have a color gradient over a message but also if [item] is in the message, that's replaced by a placeholder without allowing any other tags to work in the message

[cloud vapor]

1. use <item>
2. create a mm instance with the tags you don't want a player to use
3. run stripTags with that instance on the input string
4. create a new mm instance with the item tag
5. prepend your input string with the gradient tag you want
6. using the instance in step 4, use that to make your resulting component with the new input string made in step 5

[thin condor]

is it not advised to use the keys defined in papers extended Sound interface with adventure's sound api?

the docs use string literals for sound and i was thinking it would be nicer to use the ones provided by this interface as a sort of enum? is there any supported way of doing that or is that discouraged for some reason

[sterile rampart]

the adventure docs do not assume the existence of paper, but ya it implements Sound.Type for a reason

[thin condor]

idk why i didn't think of that that makes a lot of sense

ty

[grave smelt]

is there a default mm instance that has no tags at all that I have to add to?

[rare sage]

you can build your own instance, MiniMessage.builder().tags(TagResolver.resolver(... base tags)).build()

[grave smelt]

So here's the thing, this works but not really. my tag in the middle is computed as part of the gradient when it shouldn't be so I lose half of the gradient to the tag.

[01:42:42 INFO]: <gradient:#00ffa6:#fc4efc>adsjkfhjkahdsfjkh <2b055d2b-aef0-4ab0-8de4-c73b95328070> nasdfnfjasdnjk
[01:42:56 INFO]: <gradient:#00ffa6:#fc4efc>fajhfjdbjasbnfjhkbakjsldfn

[wet kestrel]

How about a healthy </gradient

[grave smelt]

same thing
[01:58:09 INFO]: <gradient:#00ffa6:#fc4efc>dasfjfdkjnafdsjk <6da0109f-d659-4cba-adc7-640ba4e2ed53> fasdsafdasdfas</gradient>

[wet kestrel]

You were saying your tag shouldn't be computed, so put the </ before the tag.

[grave smelt]

I don't understand what you mean here

[wet kestrel]

You did <gradient>words<tag></gradient> and I'm suggesting <gradient>words</gradient><tag> if you don't want <tag> counted.

[grave smelt]

I understand, of course I could do that but the important part here is that the tag is a placeholder that the user puts in the middle of their text

I just don't want the tag to take from the gradient

[wet kestrel]

You have users manually entering entire UUIDs into chat? Sorry, I didn't see that coming 😆
I'll let someone more skilled in minimessage help further

[grave smelt]

That's not what I'm saying

a user says [item] in chat (or whatever, it's a regexp pattern), I replace that with <$randomUuid> so that it's picked up by my custom tag handler

in chat the person wrote ...text...[item]...text...

[wet kestrel]

Okay, and then you're slapping a <gradient> on the front of it?

[grave smelt]

yeah

that's their chat color

<chatcolor>.....text.....[item]......text.....

[wet kestrel]

Intense! Thanks for the clarification. Outside of the below thought, I'm out of ideas. Someone else may have a better idea.

I would remove the tag, process to minimessage, walk the tree until I find the place the tag belongs and insert its component. I don't have an answer for how because I've never done that but that seems like a logical strategy for it. May not be the best one though.

Heck, maybe not fully remove but put in some character players can't send, and use that to find it?

[grave smelt]

the solution is just to fix the bug in Minimessage's gradient parser code to ignore tags

[wet kestrel]

If you PR that, awesome!

[naive sphinx]

If the issue is associated with minimessages, do I need to reproduce it on paper?

[robust wharf]

I mean, thats the easiest

but web viewer will work the same for us

[naive sphinx]

The issue is:
A few months back I said that minimessages performance was horrible. At that time I used legacy colors, new colors and a lot of gradients. I was notified that gradients with legacy colors can be the issue.
We now removed all of legacy colors from the server, and on that server reduced the gradients to a minimum. Texts that update each second don't have gradients.
There is a bit over 4000 text displays, by my calculations up to 1500 updates each second.

1 player in a survival server with almost everything vanilla, normal mob spawning, playing as usual, is using up to 1ms of tick
1 player just standing in a boxpvp server with all those text displays, with no other active entities, no random tick time, is using up to 5ms of tick
Will send a spark report a bit later

https://spark.lucko.me/LFqQ95ovyR

I don't understand why, but the spark report is messed up

[cloud vapor]

No it isn't, this works fine if you're doing tags correctly - please send some example code and I'll tell you how to fix it

[cloud vapor]

I mean, I can't see any performance issues there beyond "why are you updating 4000 text displays when only one user is online"

[naive sphinx]

1500*

Timers, tops, etc

[sand drift]

I mean, doing stuff takes time

[naive sphinx]

Will need to double check the tops, maybe update each minute

But mine timers needs to be each second

[sand drift]

Not really aware of there being any real performance issues when MM is used in a sensible manner, any parser will bog down, especially when dealing with anything more complex than basic string manip

[cloud vapor]

We're also only looking at 38 ticks there over 6 minutes

You're likely just picking up ticks where gc hit in

[naive sphinx]

I'm afraid that with increased player count the issue will be way worse

More text displays

[cloud vapor]

1500 pieces of information updating every second for one player online is still a bit insane

I'd suggest not doing that

[cloud vapor]

but again, this ^

[naive sphinx]

Limiting it via distance in such a small place isn't logical, especially when with more players the distance will be irrelevant as players will be more spread out.
But yea, need to check with more players. Maybe processing all that text will always take 5ms, and player count won't have that much of an impact

[cloud vapor]

Well yea, more players isn't going to make a static amount of work take longer

But again, 38 ticks

[sand drift]

I misread part of the report, just under 2/3s of the the report is literally converting it to be stored on the entity

i.e. memory churn and potentially a "you're expecting magic from overly complex components" territory

[cloud vapor]

You can see your mspt avg is 1.4ms so it's not taking 5ms for the majority of the time

[sand drift]

That too, yea, you'd basically need to benchmark it using something other than a sampler if you wanted to see how much time it was actually taking to do stuff

[naive sphinx]

Got it

Okay

[fathom mulch]

how can i make a custom tag resolver that replaces the text inside of it with different text?

[prisma mason]

Isnt that just Placeholder.unparsed or Placeholder.component?

[sweet hornet]

You're most likely going to need to implement Modifying, but can you tell more about your use case please?

[fathom mulch]

im trying to make a custom tag that takes the contents and turns them into the small caps font

as far as i know that font is not built in

[sweet hornet]

I did a prototype for this few times ago, and yes you have to create a tag implementing Modifying, and paye extra care about the fact that not all components are text (thus, you cannot modify everything)

[fathom mulch]

thank you!

[sweet hornet]

I suggest you first try to do it "simple" like by just using String#toUpperCase() for example, so you can test your text modification logic, and then you can improve it by adding the small caps conversion logic

Because Modifying are powerful, but also quite surprising 😄

(Just to save you some time, but you'll see it yourself, you need to clear the children of the component if you don't modify it, for example, because they'll get re-added by the implementation)

[fathom mulch]

not sure what you mean by most of this

i'll see what i can do, thank you

[obsidian nimbus]

https://pastes.dev/t5z8cvYn2O

i got you covered

oh im a bit lae

[fathom mulch]

Brotha you ain't late

[zealous meadow]

https://docs.advntr.dev/minimessage/format.html#gradient

I think this example code maybe doesn't match with example image. is my though right?

[fathom mulch]

by the way for reference, this is how i did it on my own https://pastes.dev/jxyvcsuZCR

yours is def better, i dont think mine preserves the style

many thanks!

[regal frigate]

Hey, is there a reason why source removing from the global translator returns false?

final GlobalTranslator globalTranslator = GlobalTranslator.translator();
if (translationStore != null) {
    plugin.getSLF4JLogger().warn("Removing existing translation store");
    final boolean removed = globalTranslator.removeSource(translationStore);
    plugin.getSLF4JLogger().warn("Removed: {}", removed);
}

Console output:

[12:45:13 WARN]: [MyPlugin] Removing existing translation store
[12:45:13 WARN]: [MyPlugin] Removed: false

It is definitely registered, but it's populated with translations after being registered in the GlobalTranslator.

Full method:

// ...
private MiniMessageTranslationStore translationStore;
// ...
private void initializeTranslationStore() {
    final GlobalTranslator globalTranslator = GlobalTranslator.translator();
    if (translationStore != null) {
        plugin.getSLF4JLogger().warn("Removing existing translation store");
        final boolean removed = globalTranslator.removeSource(translationStore);
        plugin.getSLF4JLogger().warn("Removed: {}", removed);
    }

    translationStore = MiniMessageTranslationStore.create(TRANSLATIONS_STORE_KEY);
    translationStore.defaultLocale(Locale.ENGLISH);

    plugin.getSLF4JLogger().warn("Adding new translation store");
    final boolean added = globalTranslator.addSource(translationStore);
    plugin.getSLF4JLogger().warn("Added: {}", added);
}
// ...

[regal frigate]

Well, it seems that if the TranslationStore was mutated after registration, it can no longer be unregistered using the same reference. Even though it has a key that uniquely identifies it.

[cloud vapor]

that's really bizarre

what an odd bug, thanks for reporting!

[tough peak]

Hey what is the use of ClickEvent.custom() and how can I use it?

[sweet hornet]

Recently (1.21.5 i think?), Mojang added a new "custom" click event, where the client just sends a packet with arbitrary NBT data

So this click event is just the implementation of that new feature, allowing you to use it

[tough peak]

so not much useful eh?

[little elbow]

its main purpose is dialog form feedbacks

[tough peak]

oh

ok then

[little elbow]

also, the callback system in paper is implemented with it now, not with a command click event wrapper

[tough peak]

thought it used to call an event or something

[little elbow]

Paper has an event that does fire when the custom payload is received, yes

[nocturne marsh]

Hey,
is there a way to convert a com.mojang.brigadier.Message to a Component ?

[prisma mason]

MessageComponentSerializer I think

But that's only on paper IIRC

[nocturne marsh]

thanks that's what I was looking for
I use paper

[devout sage]

why is EntityType.getTranslationKey() deprecated? What should I use instead?

[quaint monolith]

EntityType#translationKey()

[vivid bobcat]

how can I remove MiniMessage default tag resolvers? is it possible?

[rare sage]

you can build your own instance, MiniMessage.builder().tags(TagResolver.resolver(... base tags)).build()

[vivid bobcat]

thanks I'll try this out

[wet kestrel]

Please do not misuse the help channels.

[devout sage]

Is there convenience api for generating reference components to entities/players the same way vanilla death messages and chat messages do? for example:

"<insert:hjk321><click:suggest_command:'/tell hjk321 '><hover:show_entity:player:f90efb8f-6362-43f0-ab7c-6be7f8ec242b:'hjk321'>hjk321"

"<insert:27348008-73cd-4ffd-8791-6d7d5d7d1fa5><hover:show_entity:skeleton:27348008-73cd-4ffd-8791-6d7d5d7d1fa5:'<lang:entity.minecraft.skeleton>'><lang:entity.minecraft.skeleton>"

it's easy enough to construct myself but is rather annoying to do so