docs-website

[pliant canopy]

And I saw many people ask questions about stuff that is inside the docs, but could well be missed if you don't go into the correct section

So I had just assumed, that if those people didn't find it, there isn't a search bar

[spice temple]

I wasnt sure, it could mean we should improve the search

[pliant canopy]

I mean, the search itself works fine, I tested it out. I just hadn't seen this bar at the top-right

Never looked too much in that top-right direction

The paper docs needs their own docs

[still apex]

it should definitely be improved though

[vocal halo]

Maybe put it next to „misc“ then it is somewhere where people look at frequently

[trail jasper]

or just make it bigger
i think double of the current size should be enough to not miss it

[pliant canopy]

or maybe put it above the list at the right

[vocal halo]

Eh that would imply that the search is only for those sections

no?

[pliant canopy]

What if we make it look like this?

I think that'd be fine

Or we can put it at the bottom

This is how bottom would look like

I feel like that'd even better

[still apex]

no

[small harbor]

if i want to find a search bar i would expect it to be at the top of the page on the navbar

usually on the right hand side or center top

putting it in the sidebar seems odd

[pliant canopy]

I kinda like how it looks

[spice temple]

is people not finding the search bar actually an issue?

[small harbor]

i don't hate the one where it's in the top sidebar but it is still an unusual spot for a search bar

not to mention it does not make much sense to put it there

[pliant canopy]

(It's for me 😭)

[spice temple]

if so, I propose this

switching the nav bar around to have search on left and links on the right

I didnt even see that we have links on both sides right now, thats super funky

[pliant canopy]

It's just that the docs are very left-heavy

so you usually look at the lef side

[vocal halo]

The thing is there isn’t really any information on the right side other than socials and a version drop down. I feel like placing it next to misc makes more sense as it’s Center and at a spot where people might look at more frequently.

[pliant canopy]

Precisely my point ^^

[vocal halo]

And as strokkur said, with large displays it may be difficult to spot the search bar when the majority of content is on the left side of the screen.

[pliant canopy]

I like this

Time for another PR

Well, before that, maybe ask @drowsy zinc or @neon epoch for opinions

[pliant canopy]

Opinion on this?

As the new style for docs

so that search bar is on left side

[neon epoch]

Ummm

It’s a bold change

[pliant canopy]

Yeah but all content on left, except search bar right is kinda bad

[neon epoch]

Mhm

Idk how I feel about it

Maybe, I think you could just change the colour of the search bar tho to have the same effect

[pliant canopy]

that sounds awful

[drowsy zinc]

mhm I don't know

[trail jasper]

the nav bar is already really crammed on the top left
i think it should stay on the right where it is right now
but making it bigger would make it much more obvious

[eager plover]

it's all about being able to take advantage of visual trickery and aesthetics to make stuff feel more fluent and pleasing to the eyes, and often, making it easier to read; the context between the text and the bg is probably not as great as it should be, for ex

[lean venture]

Did someone say design?

[eager plover]

#gifs-and-memes message - I feel a slight tange of relevance

[lean venture]

just realized I was reading 4 day old chats

[sage sage]

mrbeast miniatures pov

[ocean void]

@sage sage Gifs and memes are best in #gifs-and-memes

[sage sage]

yeah sry for that

[pliant canopy]

Who pinged me and why

[ocean void]

It was a random gif from the user above, which I deleted, so ended up a ghost ping. Sorry.

[pliant canopy]

[vocal halo]

I remember there have been talks to change the DO NOT REPORT THIS TO PAPER part of a thread dump. But what if we change up the message to link to a docs page explaining the dump and provide simple steps to try to fix the problem. We could link to the discord as well.

[spice temple]

Yeah

[hard quartz]

git push force old friend

[neon atlas]

mr ollie or scorp https://github.com/PaperMC/docs/pull/542

[still apex]

no

[neon atlas]

[outer elk]

📖 Learn Java

To create plugins, learning Java beforehand is like learning how to write before creating a novel - it's an essential part of the process.

There is no shortcut to learning programming, just as there is no single, definitive way to approach it. Everyone learns differently. One of the best ways to learn is by doing and actively applying what you're learning as you go.

Learning Resources:

• JetBrains Academy
• University of Helsinki's Open Course
• The Official Oracle Java Tutorials
• Udemy Course by Tim Buchalka (Paid)
• GitHub Education Pack
• List of traditional book media

The best time to start learning was yesterday, the second best time is NOW!

[outer elk]

Where did paper.yml go?

In 1.19, paper.yml has been split into two files, both in the config directory. In paper-global.yml you will find configuration that changes behavior of the whole server, and in paper-world-defaults you will find configuration that can be overridden on a per world basis. See https://docs.papermc.io/paper/per-world-configuration for more information on overrides. The function of server.properties, bukkit.yml, and spigot.yml remains unchanged for this time.

[pliant canopy]

Can we not turn this into a channel for bot spamming?

That's what #bot-spam is for

[spice temple]

Can we leave moderation to mods?

[brazen moon]

noticed you removed entity-activation-range.villagers (spigot.yml) and resolve-selectors-in-books (paper-global.yml) is it intended, Noah in https://github.com/PaperMC/docs/commit/7a5c13ca250449036cb3d2a22c20a3449bc90b15

[vocal halo]

No it’s not? Strange

Was it removed for a short time? I remember some values were removed.

[grand dew]

on https://docs.papermc.io/paper/dev/command-api/misc/basic-command you can replace the getLifecyleManager stuff with https://jd.papermc.io/paper/1.21.4/org/bukkit/plugin/java/JavaPlugin.html#registerCommand(java.lang.String,io.papermc.paper.command.brigadier.BasicCommand) for the basic commands

[pliant canopy]

Don't we do that?

Ah nvm

That's JavaPlugin

Interesting, didn't know that's a thing

[grand dew]

was pr'd like last week i think

was in https://github.com/PaperMC/Paper/commit/b386a8f5273237d2dbe882832ea6c2e3ac0ee0b1

[pliant canopy]

Ah cool

I might just keep it like this for now

Not everyone might have the new api yet

because of maven caching and stuff

[brazen moon]

not those, i will update the docs in that case

[vocal halo]

okay. Sorry for that.

[last bear]

Is it intentional that the docs repo doesn't have auto branch deletion on merge enabled?

[sharp pivot]

doubt it

[spice temple]

Never noticed that, refined GitHub does that for me on any repo, lol

[vocal halo]

real

[mighty dirge]

Should this be updated because the permission is no longer true by default? Or should we wait because it is only supposed to be a temporary fix?

[neon atlas]

Should probably be updated, I don't think we'll revert that even once a packet limiter is pulled

[neon atlas]

Finally 🙏

[mental sleet]

(67d03fb76ed5010734cf1be5) // @empty nova (@_oskoskosk_ / 303257590572253184) has been banned by @neon atlas (202850073812402177)
Reason: Quick-banned for sending a message in #docs-website

[pliant canopy]

https://github.com/PaperMC/docs/pull/549 to be fair though, I was also confused for a very very short while when I first read through that. Perhaps mentioning the actual static import would be useful. Either inside the code block or right outside of that

[eager plover]

I mean, it's generally fairly clear if you understand java and read it

I'm not against adding more stuff for the sake of confusion, but, the general issue here is knowing where to set the bar of expectations, otherwise, rabbit holes...

[vocal halo]

Yeah I don’t think that change is needed. The comment also describes that static imports can be used. So adding the class there doesn’t make sense. And if you do that you can just delete the whole example because there is the same code above that one.

[raw breach]

ollie or scorp quick approval on https://github.com/PaperMC/docs/pull/550 would be nice if youre around

[neon atlas]

I'd add some note regarding the "fixes attribute exploit" or whatever it was called

[hard quartz]

i get a question (i tell ask here but no.. the ignore that thing xd) about this setting, talks about the chunk generation but also include the chunk load not?

[neon epoch]

Yeah @raw breach what he said ^

[raw breach]

You mean on the page where we already talk about some of the exploits more in debt?

[neon atlas]

No just in the thing

"Needs to be disabled in order to allow attribute swapping exploits"

or something

[raw breach]

Yeah already added something

Unless github Web fucked it up and didn't actually commit it

[neon atlas]

haven't checked 😅

Yea sounds good imo 👍 can wait for @neon epoch tho, given he assigned himself

[neon epoch]

Classic mobile GitHub moment

Thanks tamion

[raw breach]

Thanks. Yeah i saw kenny merged the paper pr but not the docs one and that was triggering me lol

[fair river]

Kenny bad. bad kenny.

[last bear]

Well you see, I knew that the docs weren't really enough

but didn't care for it in the Paper PR

[mental sleet]

(67d2e6e06ed5010734cf1be9) // @crimson tangle (@yajan / 579908961080311818) has been banned by @neon atlas (202850073812402177)
Reason: Quick-banned for sending a message in #docs-website

(67d2e6e46ed5010734cf1bea) // @crimson tangle (@yajan / 579908961080311818) has been banned by @spice temple (134340832093405184)
Reason: steam scam

[neon atlas]

[spice temple]

Smh

[pliant canopy]

Bro got super-banned

[echo canyon]

I think the adventure docs would fit in with what ollie started here https://github.com/PaperMC/docs/pull/346

essentially making each project have it's own docs site that you could easily switch between. Rarely do you need a single menu bar/ side bar that has docs across multiple projects.

My thought is that (1) would be a dropdown with the icon/name of all the projects that are part of PaperMC

then the top dropdowns would be switched from Paper, Folia, Velocity, etc., to the categorie within each project

because also, you rarely need docs across multiple categories within a project. Making the side navigation all admin or dev is useful because its 1 less level of nesting

so that main dropdown would have like Paper, Velocity, Folia, Adventure and the correct icons to go along with that.

[opal flare]

it'd be nice to look into switching away from docasaurus if possible

[pliant canopy]

Yeah I like that idea a lot

[pliant canopy]

I believe scorp was looking into that? Not sure what the outcome ended up as though

[neon epoch]

I mean tldr is it’s a lot of work to change from docusaurus to something else. I got a decent way through swizzling that functionality but uni has absorbed my life nowadays

[vocal halo]

What would be the alternative?

[eager plover]

brb, migrating to writerside

[vocal halo]

😭😭

[eager plover]

I think that the notion that we've come across is that pretty much all of the solutions that exist kinda suck, and so, idk what we'd be looking to migrate to

[vocal halo]

I guess paper needs its own documentation solution

[mighty dirge]

Seeing as Paper has environment variables listed at the bottom of the system properties page (Only one seems to be listed), should something like that be added to Velocity as well?

[opal flare]

Starlight, maybe

[lilac edge]

^ esp since the main website is being transitioned to astro

[lean venture]

Wait what?

Astro?!

[spice temple]

The branch is pretty dead

[eager plover]

Yea, idk if somebody wanted to take over that or not, it became a bit of a headache to port our stuff over more or less as-is, I had some ideas on how to mitigate it but it's all a headache

[pliant canopy]

I might look into it just for fun, but probably won't get anything presentable out of that

[lean venture]

I'd be down to migrate it to smt else, like svelte xD

[drowsy zinc]

starlight seems to be the only reasonable option right now, not many documentation generators that look good and have a comparable feature set to docusaurus

[lean venture]

What features are you looking for?

[drowsy zinc]

markdown markup extensible through custom components, a decent search, an easy way to split the site into multiple projects (i.e. paper/velocity/folia/misc/...), a tasteful design with light/dark mode, ... idk

the main headache is having the site house multiple projects separately (if you click on Paper, it should have a separate sidebar with only Paper stuff, separate search results only from Paper, ...)

I know what we have right now is kinda hacked on top of DS but thankfully I didn't have to do it

so if we're switching, I definitely don't want to hack that up on top of a different generator

[opal flare]

https://starlight-versions.vercel.app/getting-started/

[drowsy zinc]

seems like this popped up ever since I last took a look at the plugins page: https://starlight-sidebar-topics-dropdown.trueberryless.org/docs/getting-started/

[opal flare]

yeah, quite a few nice plugins

starlight is really nice from some quick testing I did a week-ish ago

[echo canyon]

what about just having like properly separate docs site, with another project in the repo that contains common components that each docs project depends on

How much "duplication" in configuration, etc. can we avoid if we just setup fully separate doc projects, then have the base url be different for each, and add the switcher in the top left.

[mighty dirge]

What about stuff like the brigadier api? Parts of that can also be applied to velocity

[drowsy zinc]

sounds like a dream to maintain

also not sure how that would work with pages publishing

[pliant canopy]

You are free to start and port it over :3

The issue what that is though that certain stuff is retrieved differently and also Velocity's Brigadier has some weird interactions, mainly when talking custom arguments

And it generally limited in a lot of ways

[pliant canopy]

That reminds me, I still have a lot of Brigadier docs left to write....

ahhh so much stuff to do

[vocal halo]

should probably go to #build-tooling-help

[glossy horizon]

Mb, wrong channel

Ty

[vocal halo]

<@&748618676189528155>

[mental sleet]

(67db2a646ed5010734cf1c0a) // @reef grail (@xnevexxx_12529 / 1251607424474026044) has been banned by @ancient quiver (1098722699116810331)
Reason: nsfw spam

[ancient quiver]

merci

[stiff sundial]

double noah

[odd harbor]

message.txt by @teal jolt: https://pastes.dev/buUmUVGVzB

[granite raptor]

since timings are disabled now, I think timings mentioned here should be changed to something else

[eager plover]

The problem here is that I don't think that we have something else

[pliant canopy]

Are those maybe somewhere in the API as like an Enum or something? Because then we could link to the jd

[eager plover]

nope

[pliant canopy]

Unfortunate

[eager plover]

that system isn't exposed in the API last I knew

[pliant canopy]

Time to expose it

(Not gonna be me though, as I don't seem to cope well with huge PRs)

[eager plover]

so there is the sensor type registry

idk if anything has a list of those values

We also calculat the name differently sooo

I'm guessing we're using the class name rather than the sensors actual name

[pliant canopy]

I mean, the value has to be validated somewhere, right?

So it'd be enough to figure that one out

[eager plover]

it's not validated, I don't think

[pliant canopy]

💀?

[eager plover]

The name is literally the class name

lowercased

so, it's not using the registry or anything

[pliant canopy]

Yeah it seems to just be a EntityType<?>, at least from what I can tell?

Aka an internal type

[eager plover]

it's entityType.sensorName

[pliant canopy]

I mean, this is what I found

[eager plover]

Yes, it's a table

2nd param is the name of the sensor

which is the missing info here

[pliant canopy]

Hm

I see

[eager plover]

which is generated from the sensors classname on init

[pliant canopy]

I see I see

Wow that's disgusting, ngl

[eager plover]

There is a registry for this, but, idk how much of a pain it would be to do lookups for this

[pliant canopy]

Hm

[eager plover]

it's not a registry, I lied

[pliant canopy]

Okay yeah it seems to be saved in MemoryModuleType<U>

But that has memory modules for basically everything - from positions, to data types, to paths, to entities

I do not think there is a particularly clean way to expose that tbh

[pliant canopy]

Actually no. That's something different

The actual sensor names seems to be, as you said previously, just the lowercase name of a class that in any way extends net.minecraft.world.entity.ai.sensing.Sensor, if I can tell correctly

Seems to be just 22. Do you think it would be enough if we were to just list them somewhere and forget about it?

Because I doubt there is a dynamic way to figure those out

[pliant canopy]

I have a proposal for a new repository: paper-cookbook.
I had seen yesterday that LuckPerms has a similar repo where they basically implemented a few common operations, put it into a repository, and then allow people to look at it to get a 'practical' understanding of a specific API. I think bigger aspects of the API, like the inventory API, or command API could really benefit from that.

Is this something that PaperMC would do or not do?

[spice temple]

Sounds like a cool idea

[lean venture]

That sounds like a VERY cool idea

Just have to make sure that examples are decently abstract

[spice temple]

and I guess they should be accompanied by guides on the docs

[lean venture]

Why do I have a feeling that chatgpt could do most of these 😭🤣

[pliant canopy]

Yeah, but it won't!

That's a personal policy

[pliant canopy]

Right. My idea was primarily influenced by me wanting to add a step-by-step hands-on tutorial on Brigadier, just in case some people prefer learning-by-doing instead of understanding-by-reading (lol). And so that basically I don't have to paste the finished code in into the docs, but instead they can look at an example repository on how it is structured, and even compile+run it if they want to!

I am also currently thinking about the layout. Perhaps the "main" branch could be just a reference for stuff and like each little docs entry can have their own branch with examples?

[spice temple]

I think folders are nicer for maintinance

[vocal halo]

yeah

a module for each topic i guess.

[pliant canopy]

Oh yeah that's a beautiful idea

Unsure about the modules, since basically none of them need to be interlinked, so I am not entirely sure you'd need that

But folder separation sounds nice

[vocal halo]

same difference really

[pliant canopy]

Hm, well, not exactly

[vocal halo]

make modules, make folders in the end it is just a folder

[pliant canopy]

The advantage of modules is that you have a "main" build.gradle.kts in which you can define stuff that's true for all sub projects. Like adding the paper-api dependency. Though I have a feeling it might be better to just declare it again in every folder so that a user basically just has to copy that small portion and it would work

I wouldn't want to introduce the need to know how to work around gradle if somebody new is going to look at it

[pliant canopy]

Yeah, ik what you mean

I was mainly talking about submodules

[vocal halo]

honestly if somebody new is looking at a repository where they find more stuff than they asked for it would probably just be overwhelming

also, new people wouldnt even know how to actuall look at code

for new people it makes sense to suggest documentation imo

[pliant canopy]

I mean, the main source of information is still documentation. The repo I suggest is mainly for one to have a working example and perhaps as a quick-reference

[vocal halo]

im pretty sure the flowchart will be as following

-> Ask in discord for example about Brig API -> Someone redirects them to the cookbook -> Person copies example into their project -> Person wants to expand the example -> person comes back to help channel

[pliant canopy]

That's precisely why we want to do this:

-> Ask in discord for exmaple about Brig API -> Link to docs page about building up an example -> The docs page links to the finished example at the end -> Person, after hopefully reading through the docs, looks at the example repo -> Hopefully understands it well enough

[vocal halo]

i like your optimism but i doubt it will be like that.

people always look for the quickest way. And if they see an "see full example here" they will click there and do not read the documentation.

[grand dew]

would def be useful for brig, when i was reading through it i got confused about how to handle multi arg comands being changed on each other not on the original comand

[vocal halo]

well, then we should improve the documentation. If the documentation is confusing i dont think an example would magicaly make it unconfusing.

[pliant canopy]

I mean, I have like an entire site documented to just command trees and argument chaining lol: https://docs.papermc.io/paper/dev/command-api/basics/command-tree

[grand dew]

icl its not the most obvious thing on some things

the location of the parenthesis is something a lot of the time i will and others gloss over

[pliant canopy]

Having colored parenthesis would be a dream in this case lmao

But I doubt that is possible with docusaurus

[grand dew]

literally tho

[pliant canopy]

Another example I could see the cookbook repo be useful is perhaps with the DataComponent api, since that one is also rather complex, at least when you hear about it the first time

I am seeing a pattern with complex api design and Owen being the maker

[lean venture]

I disagree with this honestly

That sounds like just spoonfeeding people

Which is 100% not the right way to learn

[young mural]

Hey everyone, I've been thinking of remaking the PaperMC Docs website but I think I can do it way better if I use Angular and Java, Am I able to go that route and potentially getting the website to be used, ONLY if it is widely liked by the PaperMC staff and Admins? or Am I forced to use node.js

[ocean void]

I don’t think you should spend substantial time remaking anything because it’s unlikely to get adopted by us. If you have ideas for changes you could submit PRs, though.

[young mural]

Alright sounds good

[eager plover]

Docs don't have the computational demand to justify using a full JVM, especially when we could literally just use an SSG, which we already do iirc

[spice temple]

Yeah, the docs should require a process on the server

Nor are we open to using angular

We are open to moving away from docusarus tho, it's just, idk, nor worth the effort rn

[opal flare]

https://discordapp.com/channels/289587909051416579/947676006964162611/1350254368481476629

if you want to look into anything, look into this

[vivid mantle]

vitepress 🥶

[pliant canopy]

Oh, there is new brigadier registries? Which ones are those?

Actually I can probably just look at the diff from 1.21.4 and 1.21.5 to figure it out. But if somebody just has a list ready-to-go, I wouldn't complain

[brazen moon]

it's just regular registries like entity variant / sound variant but the doc has a page for all the registries exposed to the api

https://docs.papermc.io/paper/dev/command-api/arguments/registry

the new options: enable-spam-exclusions and simplify-remote-item-matching are not documented too

[opal flare]

If moving to Astro doesn't seem feasible for the website, maybe we should just upgrade to Next.js 15

I can move the website off of Cloudflare Pages now that the legacy routes (/api, /repo) are dead.

[small crane]

What makes moving to Astro infeasible? Is it just that nobody wants to commit to it [yet] or is there some custom stuff in the current solution that isn't possible to move?

[opal flare]

The interactivity has had progress stuck for months now

https://github.com/PaperMC/website/tree/feat/astro
https://github.com/PaperMC/website/tree/feat/astro-cats-fight

https://github.com/PaperMC/website/blob/feat/astro/TODO.md

[spice temple]

basically nobody here really understands astro

[small crane]

Interactivity meaning wiring up the components?

[spice temple]

the idea seems great but we actually have interactivity and then everything seems to fall appart

[small crane]

I can take a crack at it later

[opal flare]

Basically making all the old react stuff work - we're fine with pulling react in for the pages that need it (like the download explorer), or svelte

[small crane]

Is it all in there right now, just not 'working'? I need to finish some stuff at work then I can try to look at it

[eager plover]

There is a branch, and then there was my branch that I pushed a thing or two to

[opal flare]

it's been a few months since i last took a look; I think a good portion of it is there, but I think some still needs porting

[small crane]

Alright I'll take a look in a few and see what's going on

[opal flare]

will happily take any help we can get with finishing this migration

as an idea of how out of date it is, too:

┌────────────────────────┬─────────┬────────┐
│ Package                │ Current │ Latest │
├────────────────────────┼─────────┼────────┤
│ @astrojs/react         │ 3.6.3   │ 4.2.3  │
├────────────────────────┼─────────┼────────┤
│ @astrojs/tailwind      │ 5.1.5   │ 6.0.2  │
├────────────────────────┼─────────┼────────┤
│ @types/react (dev)     │ 18.3.20 │ 19.1.0 │
├────────────────────────┼─────────┼────────┤
│ @types/react-dom (dev) │ 18.3.6  │ 19.1.1 │
├────────────────────────┼─────────┼────────┤
│ astro                  │ 4.16.18 │ 5.5.6  │
├────────────────────────┼─────────┼────────┤
│ react                  │ 18.3.1  │ 19.1.0 │
├────────────────────────┼─────────┼────────┤
│ react-dom              │ 18.3.1  │ 19.1.0 │
└────────────────────────┴─────────┴────────┘

[small crane]

Sounds like fun

[lean venture]

if it were svelte I'd be here in a heartbeat xD

[small crane]

I mean, you can use Svelte in Astro.

That's actually the plan here, as far as I can tell.

Anyway, I just forked and I'm taking a look. Let's see what the damage is.

[eager plover]

I remember us pulling svlete in for interactivety because somebody glares at mini basically said no to just pulling react in

[small crane]

Everytime a new version group comes out, do you guys rebuild the website for the software pages to update the javadoc link? Or is it dynamic? If dynamic, I think we need to use interactivity there since astro is only going to make the web requests at build time. (Unless you switch to SSR)

[neon epoch]

It’s dynamic

[spice temple]

It would be fine to trigger a rebuild for that tho

But ye, don't mind it being client side

[last bear]

The downloads site already requires rebuilding

[small crane]

I have no idea how next getStaticProps works then (current), because usually that requires a rebuild unless the current site isn't SSG.

[eager plover]

is that the astro branch or?

apparently we're using ISR?

revalidate: 600

[small crane]

https://github.com/PaperMC/website/pull/132

wife is angry that im not getting ready for bed so that's all i can do tonight

[eager plover]

gdi simple

o/

[small crane]

Oh ok it runs every now and then. Makes sense then. Don’t think Astro has that so will need clientside

[opal flare]

merged this one (as your description mentioned)

[lean venture]

b- but I want ONLY svelte

[pliant canopy]

@brazen moon in case you got some spare time, would you mind looking over feat: Document custom arguments #553 to make sure I didn't write anything stupid?

The added page is this one: https://strokkur424-feat-custom-argu.papermc-docs.pages.dev/paper/dev/command-api/basics/custom-arguments

Also, I still need to finally add the CommandSyntaxException page. I've used those all over the page, but they aren't documented anywhere at all

[brazen moon]

probably not until middle of the month sorry

[pliant canopy]

No worries! I took like 3 months until I came around to finally write that one, so it is definitely not like I am on a time schedule

[small crane]

@opal flare:

What is the goal for the site? Client-side web requests or SSR? If client-side, we can do some skeletons while the requests finish loading. If server-side, then we need to enable SSR so the web requests can run on the server or keep as is and just turn off pre-rendering for the pages that use fetch. Right now fetch() in astro only runs at build time unless prerendering is turned off.

If you have a direction you want to take, let me know and I'll keep working on this.

I think clientside is a fine way to go personally. Most of the site can be pre-rendered then, and only a few dynamic clientside islands would be needed to load the dynamic content.

But SSR is good too, SEO is best in that case, but I think it would probably lead to higher workers cost.

*Unfortunately we can't use ISR like the current Next site since AFAIK Astro doesn't have that.

[opal flare]

If we went the Client-side route would we need to rebuild the site for new builds of our software (paper, velocity, etc)?

[small crane]

No, it would automatically update. The only downside there is that dynamic content won't show up in search engines (specific version groups, builds, etc). But that's probably not important. And then there would be skeleton loaders that show up while the request runs

[opal flare]

Do you think SSR would be better? We’re currently hosting on Cloudflare Pages, but if the site is Docker-ized we can self-host it and avoid Worker costs.

[small crane]

The biggest benefit of SSR is that the fetched dynamic data actually ends up in markup, which can enable search engines to show that content in search results. A minor benefit is usability - the user won't see loading UI, they will just get the content at the cost of a slightly longer page load. In this case, I don't know if the SEO benefits are really that great here since the dynamic content we're fetching is version groups, builds, etc. So probably not a huge boon for Paper, it's more useful if the content is actually semantic (blog posts and the like).

So we could really go either way. Looking away from SEO, the question here is:

• Show loading UI while fetching version/build info? : Clientside
• Don't want loading UI, when page loads the information should be there already : SSR

If you have the processing power to spare, SSR will be simpler codewise since we won't have to worry about handling loading state in the components. That's one other thing to consider as well.

[opal flare]

If we went with Client-side today, would it be very difficult to migrate to SSR if we need to?

[small crane]

You'd have to get rid of the loading state components and basically strip away some conditionals in the templates. So not really.

[opal flare]

I’m thinking Client-side might be the way to go for now, then, unless you highly recommend SSR

[small crane]

I have no real opinion either way. Both are fine. I'll go with Clientside for now.

[spice temple]

What's even the point of Astro if they can't do incremental stuff and we ship a full fronted framework anyways?

That's what I meant with I don't get Astro

[small crane]

Most of the website will be a pre-rendered static site, with only small parts of it being clientside code (Svelte/React etc whatever). That's the biggest difference. With Next static generation, your whole app is React. With Next SSR, it's still React but a good portion of it is server rendered.

With Astro, your Astro components get compiled down to HTML/CSS and sent to the user. Only the small parts that are actually dynamic will be sent down as a JS app in whatever framework.

So it basically sends a hybrid page, which is different in a unique kind of way.

That's their sales pitch anyway.

[still apex]

we should switch to wordpress

[spice temple]

At work I am building a fun thing, maybe I can replicate that some day because God knows they will not let me open source that as a framework:
It's a "normal" Vue SSR app (using vite at dev time for hmr, express as a web server, maybe I'll replace that). Cool part is that during SSR, I first look at all modules that were used for the request and inline that css, dynamically, and then look at all the Vue components and which of them are interactive and only send the JS code and state for those, mounting many small Vue apps around all the components that are client interactive on the client, kinda like Astro but without some of the limitations

It's not built for isg but should be easily doable

And it's like 200loc

And it will make my website so much faster and lighter than nuxt, lmao

[small crane]

yeah that is pretty much what astro does, it makes little "islands" of apps, but otherwise everything is html/css

Without frontend UI frameworks, Astro is essentially just an HTML/CSS framework. It takes all your pretty components and smashes them into static HTML/CSS. When you do a network call in an Astro component, it just runs that call at build time. That's great when the stuff being requested is also rather static. But when it's dynamic like what we have here, it's not helpful. So you either need to enable Astro's server-rendered mode, which basically just acts like a preprocessor and runs the script on each request (which would include the fetch) and then returns the static HTML/CSS, or you make a clientside island and have it run your request then.

[lean venture]

I personally use SSR almost everywhere I can. The biggest deal being SEO as stated. And it usually improves loading times too.

[spice temple]

We want ISG

Or we would throw it on our hetzner servers

Ain't no wait I am paying cloudflare for ssr

[small crane]

I’m actually able to make use of web components here for some of the stuff since it’s really just data fetching. No framework bloat (yet)

Probably will need for build explorer later though

[small crane]

https://github.com/PaperMC/website/pull/133

Goodnight peeps

[still apex]

domain and such need to stay as .dev for now

iirc that branch is published on there and we don’t want that indexed

hence the TODO

[small crane]

Alright will edit it back later

[opal flare]

It doesn’t deploy as far as I know

[small crane]

I removed the domain change just to be safe, will worry about changing it as the final thing before finishing the main PR I suppose.

[small crane]

Currently applying placeholders where hangar requests are considered as there is a CORS issue (see #hangar-contrib) but will go back and address it later once we figure that out. Already made a Hangar PR that I think should resolve it.

[small crane]

Alright. The software download page has some very arcane code that is making my head hurt so there will be some delay on this, but I'll figure it out eventually

[vocal halo]

The download explorer broke everyone I think.

[spice temple]

Feel free to throw it away and redo if that's easier

[small crane]

Might end up doing that

[mighty dirge]

How does the software version thing exactly work?
Because the Velocity compat page still shows 1.21.4 as max version.
-# https://docs.papermc.io/velocity/server-compatibility#compatible-game-versions

[pliant canopy]

I mean, there is no released 1.21.5 build

It is internally handled by the <SoftwareVersion/> tag, which relies on the latest released Paper version afaik

Yep, it uses the Paper version. Looked it up

[mighty dirge]

Velocity already has 1.21.5 support.
bruh, ig this will work better in the future with the improved paper tooling but this is a bit annoying.

[small crane]

This is actually just because they're missing specifying velocity as the project to the SoftwareVersion component. The default prop when not specified is Paper so it returns the latest Paper version. If they specify project=velocity (or whatever the docs-site specific way to reference velocity is), it will work.

Oh just kidding, nope. It needs to use paper because velocity versions don't track with MC versions. So just kidding.

[pliant canopy]

lmao

I mean, once the tooling is figured out, the next versions should be very fast

So this is just temporary ig

[pliant canopy]

@neon epoch do you think putting the note above the ice cream example and also putting them into the warning block would be better?

I'd still like to have some sort of header above it just to be able to paste a link somewhere which directly goes there. But now it is wrapped up more neatly

[vocal halo]

i think the main point of ollie was that the header was "note: ..." and that if its a note you can put it into an admonition

if the header is "error handling during the suggestions phase" it can probably stay as normal body

[neon epoch]

Yeah what Noah said. Being able to link to it is a valid reason to not do that, I just thought it could look better in an admonition

Personal preference tho

[pliant canopy]

I suppose I will just keep it like this then

I pushed the last of the changes. Should be mergable now

[south thunder]

Er, what do I do if I'm making a PR to the Docs but the spell checker yells at me because Leaf's code is written in canadian english?
I include a stack trace that contains "prioriti__s__ed"
[18:23:38 WARN]: at ca.spottedleaf.concurrentutil.executor.standard.PrioritisedThreadedTaskQueue$PrioritisedTask.executeInternal(PrioritisedThreadedTaskQueue.java:351)

[drowsy zinc]

add PrioritisedThreadedTaskQueue to .typos.toml into the default.extend-identifiers section

[pliant canopy]

Lmao oops

Accidental concurrent fix

[neon atlas]

however produced the right image wins

[hard quartz]

rock paper scissors for what PR keep? xd

[pliant canopy]

😭

Let's play one round of hypixel duels 💀

Whoever wins, wins the PR

[hard quartz]

i dont have context for how are you in pvp then decline... better @neon atlas throw a coin

[pliant canopy]

Yeah fair

[neon atlas]

I like the right side more tho

I think a warning is too much

[pliant canopy]

Me is heads and doc is tails?

(but userdev page is mine to maintain )

(I have done the previous 3 prs on it)

-# (That means it is mine!!)

[neon atlas]

Sometimes you gotta let go

[pliant canopy]

fr though, it's not mine to decide. If you really think the other one is better, feel free to close mine

I am just messing with you lol

[hard quartz]

looks like the same than the coin then just wait what aproach is good

[neon atlas]

I'll flip one in a couple hours when I am done writing thesis stuff for today 👍 thank you both already

[vocal halo]

Right one

Looks better. Warning seems overkill imo

And doc was first in theory

[neon epoch]

Shit looks good @drowsy zinc

[opal flare]

looking good @drowsy zinc

[minor marlin]

how to download lastest version of paper for server? (by url pls)

[outer elk]

https://papermc.io/downloads

[scenic gull]

But please ask in the correct channel in the future. This channel is for our documentation.

[neon epoch]

https://github.com/PaperMC/docs/pull/559/files#diff-8500ec99102af548cacc7ea275993818e2811abe022f08ef9d64baaa849e6677R14 any lurkers have any opinions on these pieces of software

I’m not all for recommending specific backup software

[drowsy zinc]

borg is pretty ubiquitous among backup softwares, don't know what kopia is

[hard quartz]

i read about kopia but was a random post in reddit when search options for backups

[south thunder]

Basically, a Windows alternative

Pretty much the only established, open source alternative I know of that also supports incremental baackups

[pliant canopy]

The best backup alternative is github/git

Prove me wrong

[opal flare]

@drowsy zinc is there a good reason to do the jd: url hack instead of using a custom component?

[drowsy zinc]

it's way easier to type and you don't have to be in a MDX file to use it

it's not really a hack, it's just traversing the AST and replacing the values

[pliant canopy]

@drowsy zinc I also just noticed something. You changing basically all of the files means that the "last edited by" will show you on every single page, won't it?

[opal flare]

It just feels kinda hacky to me since it’s abusing the url format a bit - a component would allow us better control I think vs rewriting everything again?

It might be worth moving to a component now during the initial move instead of down the line having to replace it all again

[pliant canopy]

Tbh, I think I am on scorps side here. It is way cleaner this way

At least I personally prefer it

Shorter and more expressive, at least that's what I think

[`PlayerMoveEvent`](jd:paper:org.bukkit.event.player.PlayerMoveEvent)
// or
<Javadoc name="org.bukkit.event.player.PlayerMoveEvent">`PlayerMoveEvent`</Javadoc>

[opal flare]

I’m just not a huge fan of it. It feels like a slight downgrade - going from an actual component with named arguments to string splitting. Less control when it comes to the "rendered” look unless you manually go and update every single use of it, when it comes to styling etc.

[pliant canopy]

I mean, where do you need to change the style of your jd link

And what if there is just both? So you can do the string splitter if you just want simplicity, or you can use the component if you actually have to change its style

[opal flare]

It just feels like this is something that a component should still be doing, like it was on the docasaurus branch. It isn’t as friendly - appending more and more separated with a : vs named arguments to be clear of what something is. Since you can technically do something like <PaperJavadoc name=“org.bukkit.event.player.PlayerMoveEvent"/> to have the proper project/module applied, automatically give it a name unless manually specified, etc

vs manually specifying modules/projects/etc in each location

[pliant canopy]

A thing which I added when I made my own jd components was add an id field which basically predefined that whole path

So you could just write <JavaDocs id="PlayerMoveEvent" />

But yeah, I get what you mean. You can extend it easier

[opal flare]

But this is why I would prefer to use components instead of the url-like path @drowsy zinc

[opal flare]

Yeah, exactly

[pliant canopy]

Another something that I did in my jd component was to infer the project by the package. Is that something that could also be of interest? Because that could remove the need to specify the project directly, at least for most stuff

[drowsy zinc]

idk, I'm of the opinion that markdown should be nice to read and simple to write, mixing in mdx markup for just links feels icky to me

I could definitely do both the component and that link processor if a need for more customization arises

the format of the links can also be changed to basically any string, it's not restricted to the URL format - I just thought that was intuitive

[drowsy zinc]

there's no last edited by as of now, just an update date, but yes it would show me

[vocal halo]

Yes

That’s the case with normal docusaurus updates as well

[pliant canopy]

My ego will be severely hurt by that /s

[vocal halo]

lol

[eager plover]

mojang doesn't auto detect that stuff, you'd need to add a click event to it; #adventure-help if you need further support with adventure

[slow swift]

oh lol i missclicked the channel sorry

[wintry flint]

When will the documents be added?

[outer elk]

__There Is No ETA__

Updates to Paper do not have any sort of estimate for when they release, ever. Any and all updates will arrive when they are ready, and the only thing to do is wait for them patiently along with everyone else.

ETA means 'Estimated Time of Arrival'!

[eager plover]

The fabric docs might contain some useful information covering some of that sorta thing

[pliant canopy]

Soon™️. I added a page just a week ago, will probably do the tutorial one next

The last three are rather complex topics, so they have to be well written, which isn't all that easy

[drowsy zinc]

hey, so I'm basically done with the starlight rewrite (if I hadn't forgotten anything); if anyone has any (constructive) design, editorial, accessibility, ... concerns, I'd be happy to hear them
preview: https://papermc-refactor-starlight.papermc-docs.pages.dev/

(the Paper logo will be in the production embeds and item command converter isn't done yet, @pliant canopy is having a go at it)

[small harbor]

that looks really nice

[mighty dirge]

• I'd like to have a nav bar because right now, you need to scroll to click on a project and then scroll to get to e.g. development instead of hovering over project>click on development
• Additionally, when you scroll down in the sidebar, the project selector stays within the sidebar and does't float so it really doesn't seem convenient to switch projects atm
• I'd like to be able to click on the github/javadoc icon and directly get to that project's page

I do like the sidebar being unfolded tough, especially if you could jump to a section directly as mentioned above. I also appreciate the label(s) on the sidebar and the general compact look.

[neon epoch]

Looks super good, nice work mate!

[hard quartz]

Feel nice in mobile

[drowsy zinc]

the project dropdown has been made sticky and the padding on the hero page has been made a bit smaller, so you can see all projects on a 1080p screen without scrolling

making the social icons context dependent is going to be a bit rougher, since the javadocs need a version in the URL but it should be doable I think

[hard quartz]

take me a few seconds make the match for the "J" icon xd

[neon epoch]

LMK what you think about that scorp, might not be what you were going for :)

[drowsy zinc]

I would have wanted a different icon but starlight has a limited set of icons and I already had to patch in paper, velocity, folia and waterfall ones, so I'll leave that until the starlight team merges the custom icon PR

[hard quartz]

oh... i see that comment in the refactor.. sad currently not allow like Astro to add custom icons.

[drowsy zinc]

looks nice

I'm doing tweaks on the branch locally so please don't merge it in right now

[neon epoch]

I won’t merge it, I’ll leave it to you :)

[pliant canopy]

Yep! Should be done by the end of today

[mighty dirge]

Also just comparing
https://docs.papermc.io/velocity/dev/command-api
with
https://papermc-refactor-starlight.papermc-docs.pages.dev/velocity/dev/command-api/

Why is "The __c__ommand API" not capitalized?

[pliant canopy]

Because scorp is bad

@drowsy zinc please read through our contribution guidelines

More specifically, the style guidelines: https://github.com/PaperMC/docs/blob/main/CONTRIBUTING.md#style-guidelines

Thank you

[drowsy zinc]

they were revised to fit the design

[pliant canopy]

smh

rule bender

[mighty dirge]

Also...it would be nice to have the path thing in the new docs or the sidebar being scrolled to the section when opening a link with a sub-path

[pliant canopy]

@drowsy zinc I'd like to ask about this though? Don't you think it woudl be nicer if it were anchored at the top or something? Imo it looks weird floating and half-way obstructing the text like that

but that might honestly just be me again because my style taste is apparently very different from all of here

[mighty dirge]

I asked for that but it could be refined ||later||

[pliant canopy]

✨ later ✨

[drowsy zinc]

I would like there to be a background too but it's not feasible without overriding a ton of stuff

[pliant canopy]

Just add it directly to the sidebar component?

[drowsy zinc]

not how it works

[pliant canopy]

Yes how it works. Component overriding is a thing

Or is there something that makes this not feasable?

[drowsy zinc]

the entire sidebar component is in a padded space

the dropdown would have to be outside of that and have the same padding

[pliant canopy]

Just add a top margin to the sidebar and you are good?

Idk it seems painfully trivial to me because I haven't tried it. I trust you know what you do

[drowsy zinc]

now that I look at it, it might be possible if I override PageFrame

it needs to be here https://github.com/withastro/starlight/blob/main/packages/starlight/components/PageFrame.astro#L14

[drowsy zinc]

should be fixed

[lean venture]

I am jealous xD

migrated it so fast

[random pollen]

kenny pointed out I have wrong capitalization in other spots too (i.e. in the chrome ext manifest)

but yeah Patch Roulette Diff Viewer is the full title

[drowsy zinc]

I also seem to have missed "uses cases"

[random pollen]

hmmm I should have removed "The" also

my bad

I forgot it was there since it's on the previous line, but now that I noticed it I see why you would have thought it's not title case

I guess it's technically correct with or without 'The' it just feels like it flows better without it

not a huge deal

https://github.com/PaperMC/patch-roulette/commit/4531f6051226a77bf8924a6d2b91789cf7ea07cb there we go

https://github.com/PaperMC/docs/pull/566

should be good now, sorry about that

[fair river]

Gotta say the starlight rewrite looks fire. 🔥

[random pollen]

yeah navigating is much more intuitive with everything in the sidebar

is it just because it's a preview that going between pages is a hard browser reload? or is that intended

the old docs looks like client side nav

[drowsy zinc]

intended I think

it's all generated into static html so it goes through the site via normal links I guess

[random pollen]

that's good as a fallback when JS isn't available but imo it's kind of a downgrade generally

surely they have a way to enable client side routing

hm, I guess not(?)

it's probably fine tbh, there isn't any visual weirdness switching pages

[drowsy zinc]

astro's whole appeal is zero JS by default so I'd be surprised if starlight has that

[random pollen]

actually, I was wrong

in firefox there's no weirdness, but in edge there's some sort of weird flash of the background color

most people probably won't notice or care to be fair

ok no that was just cached vs uncached pages

ignore me

[eager plover]

it does feel kinda pretty jaring when navigating around pages

[random pollen]

there are some layout shifts yeah

like, as it loads in

let me try profiling page switches with dev tools

going between the same two pages it seems random if I get a 0 or a 1.0

the body and the search bar is what its upset about

[eager plover]

fonts are lazy loaded

css is loaded late too

[random pollen]

my browser crashed trying to load the PR diff then I remembered to use my extension

[random pollen]

weird, I see online that astro builds with vite

[eager plover]

I mean, it is defined early on inside of the file

there is some loaded at the bottom though

[random pollen]

which file is that in

[eager plover]

view-source:papermc-refactor-starlight.papermc-docs.pages.dev/velocity/faq/

[random pollen]

hm, I guess it's not lifting the CSS from the code blocks into the head

[drowsy zinc]

that can be changed https://docs.astro.build/en/reference/configuration-reference/#buildinlinestylesheets

[random pollen]

if we want to avoid layout shift then we probably want always

of course there's a tradeoff with initial paint time but I think the smoother transition is worth it

wow this is a classic firefox moment

(left is edge right is firefox)

[drowsy zinc]

looks better on my side

[random pollen]

yeah definitely better

the logo in the top left still flashes in and out, maybe need to set it to preload somehow

[drowsy zinc]

uhm https://docs.astro.build/en/guides/view-transitions/

seems like it does have client side routing

[random pollen]

oh, neat

would be kind of funny if we used that without actually using any view transitions

(they don't work in firefox btw )

[drowsy zinc]

the third-party plugin that brings view transitions to starlight really doesn't like the dropdown

it doesn't update the current project

probably because the dropdown is outside the sidebar content space

I'll poke at it more tomorrow, but it'll probably be less jank without it

[random pollen]

yeah preloading the icon is probably good enough tm

[pliant canopy]

urgh, what the hell

why did my pr screw up so hard?

(not a real question, I am just mad about it)

[raw breach]

@random pollen https://github.com/PaperMC/patch-roulette/pull/27 broke stuff. it will produce "https:github.com/PaperMC/Paper/pull/12379" which doesn't match the regex

hmm weird brave seems to be returning https: instead of https://

yeah firefox doesn't return the slashes either

[pliant canopy]

@neon epoch would you mind taking a look at https://github.com/PaperMC/docs/pull/568? I am mainly looking feedback on the UI. The actual logic is broken because the https://item-converter.papermc.io/ endpoint seems to be broken

Scorp said he'd take a look at that later, but yeah

[neon epoch]

I can have a look. Currently buried deep in 2fa shit but yeah

[neon epoch]

output = input;

UI looks nice though

have u got an example prompt i can test with

[pliant canopy]

As I said, it doesn't work

The endpoint is just broken

But for reference:

/give @a diamond_sword{Enchantments:[{lvl:1,id:mending}]}

Should turn into this:

/give @s diamond_sword[minecraft:enchantments={mending:1}]

Or something very similar to that

Just in case you want to have a try at fixing it

[neon epoch]

OH it doesn't work because of CORS

that makes more sense

@opal flare im assuming this is purposeful?

[opal flare]

iirc yes

[neon epoch]

Yeah, @pliant canopy put sensical logic in and we will review before merging :)

The UI looks good to me though

[pliant canopy]

What exactly do you mean by that? Try to write my own command upgrade logic?

[neon epoch]

Implement the request logic, the request will fail but we will review it on the grounds of logic if that makes sense

[lilac edge]

basically use dummy data until it gets merged

[south thunder]

Asking because I understand Folia is a special case -- Is there any blocker or something to keep in mind if I want to start some work in the development guide?

[neon epoch]

Not really

[ocean void]

The only real blocker is "Leaf doesn't write anything down".

But take a crack at it if you want. Will probably have to have Leaf review PRs to make sure he gives them the thumbs up.

[left egret]

I really hope our javadoc can have "since" part like this or even this("since" for every method)

[vocal halo]

There were plans to do that but it got ultimately declined since it has a relatively small benefit for the maintenance cost. Read comments on this PR to understand more about the decision. Also, there is an unofficial jd website that has the „since“ method and another tool that lists all additions between versions, https://github.com/PaperMC/Paper/pull/11771#issuecomment-2560372614

[opal flare]

@drowsy zinc any changes needed for deployment?

[drowsy zinc]

should work I think

[opal flare]

https://tenor.com/view/pull-the-lever-wrong-lever-wrong-lever-kronk-pull-the-lever-kronk-the-emperors-new-groove-gif-27333081

not sure if I want to squash this one, might just merge it in with the 46 commits or so

one oddity @drowsy zinc

[drowsy zinc]

hmm

[opal flare]

its to do with the orange background colour stuff

[drowsy zinc]

myeah seems like I didn't override it back to the proper value for light mode

I changed orange because people said it was too pale for caution admonitions

[opal flare]

ah

easy fix?

[drowsy zinc]

probably but 🛏️

[opal flare]

ah :3

left a comment on PR

[pliant canopy]

Damn. scorp getting +46 commits added to his balance

That would basically almost double his commits

[neon epoch]

It’s a lot of work 😂

[vocal halo]

Yeah

Can’t really compare it to a regular PR

[neon epoch]

Where does this image come from it’s noticeably blurry

Good work overall

I’m not really free until Saturday to give it a proper review, my surface level look looks good. I’m happy for someone else to review in lieu of myself

[hard quartz]

I remember the mention about that in the main website (?
Maybe come from that

https://assets.papermc.io/brand/papermc_combination_mark_dark.min.svg

[drowsy zinc]

safari has issues with scaling svgs I guess

nothing much I can do about that except replacing it with a png which is not an ideal solution

[spice temple]

safari being a bottom tier browser yet again

[opal flare]

looks better now @drowsy zinc

[still apex]

thanks

[pliant canopy]

@drowsy zinc in addition to the review I am about to send in, these pages from the velocity section link to the Paper one. I find that not ideal as you then get yeeted off into the Paper section and can get confused where all the sites have gone to

[drowsy zinc]

I don't think there's a good way to change that other than to duplicate the pages

those pages are inherently paper doc pages and the audiences page also references paper api

[pliant canopy]

Yeah, but it is still bad imo

Like I could get away with that in the adventure docs, since that is the same category

But here it flings you off completely

[mental sleet]

(68027ffd1df977252a68f75d) // @raven sundial (@lammy12k / 1334287928473551008) has been banned by @spice temple (134340832093405184)
Reason: steam scam

[opal flare]

🎉 https://docs.papermc.io/
big thanks to @drowsy zinc for the bulk work of the migration, and thanks to @neon epoch and @pliant canopy for their contributions too

[neon epoch]

Eyy let’s go scorp

[neon atlas]

lets go ollie for reviewing

[small crane]

woop woop. re: actual site astro migration, not abandoned just been very busy. will get back to it hopefully soon

[neon epoch]

Honestly I should get full credit for this one

[pliant canopy]

Ngl, I should get full credit for this. I was the only one who was saying constructive feedback /s

[small crane]

300% credit to go around. 200% to scorp, 50% for each of you

[hard quartz]

And for that now the docs not show hows was the last contributor

[pliant canopy]

Yeah I will say I am a bit salty about that

I have been discredited from my beloved brigadier pages 😦

[hard quartz]

Unless add a little final text for "this was started by"

[pliant canopy]

"Initially added by" is actually a crazy good idea

Though, for most pages, there is no record who initially wrote them, except if we were to check the git blame

[lean venture]

god this is sexy

what is the underlying stack/tech?

[small crane]

it's starlight (astro)

[lean venture]

oh nice migrated to astro

now I actually wanna migrate hangar to svelte

xD

[drowsy zinc]

we do also use svelte for the docs tools (i.e. https://docs.papermc.io/misc/tools/start-script-gen/)

[opal flare]

it might be possible to restore it using something based on https://scottwillsey.com/astro-last-modified/

@drowsy zinc any idea why we aren't using https://starlight.astro.build/reference/configuration/#head instead of https://github.com/PaperMC/docs/blob/main/src/components/overrides/Head.astro?

[jagged pecan]

Is this expected?

(documentation group)

[drowsy zinc]

uh as far as I can tell you can't override the OG meta to not suck with that

[drowsy zinc]

yes

[opal flare]

which part of the meta? the title?

[drowsy zinc]

anything

it looks like this by default from my testing from a while back

[opal flare]

does it just ignore the ones set when using the head config? that is odd

[pliant canopy]

May I try to give it a go?

[opal flare]

sure

[pliant canopy]

What the hell did I just write

Whatever, you understood me

[drowsy zinc]

there already is a last updated time in starlight so rather modify that

[pliant canopy]

Yes I was thinking about doing that too

[jagged pecan]

Also does it reorder stuff so that doc groups appear before pages in the listing? But not the navigation?

Rather than both alphabetical/whatever

[pliant canopy]

(Will look at that later though, because I have other stuff to attend to. Also have to rebase the adventure pr to the new main)

[drowsy zinc]

the sidebar is completely independent

the page cards are just alphabetically sorted + groups go first

[jagged pecan]

Ah are they both not auto populated from the listing?

Oh sorry I see

[drowsy zinc]

the sidebar is defined in astro.config.ts, the page cards are sibling files discovered in the collection

https://github.com/PaperMC/docs/blob/main/src/components/PageCards.astro

[lean venture]

W

[brazen moon]

the experimental flags could be expanded a bit in the sidebar and the images are really big in assets i think

[serene vault]

Any chance of moving the sidebar to the left on mobile? Right side is super unusual and not where I intuitively would look for it

[drowsy zinc]

sure

[twilit fog]

I hope this is the right channel, but when clicking the JavaDoc link in https://papermc.io/software/velocity, it redirects you to Velocity 3.3.0 JavaDoc instead of 3.4.0 (the link is set to 3.0.0)

[still apex]

knownw

[small crane]

It redirects to the current stable release, it’s a manual process to get it switched to latest as far as I know. They will get it updated soon

And the link to 3.0.0 is the link to the version group, so that’s valid

[drowsy zinc]

@neon epoch is https://github.com/PaperMC/docs/pull/346 something that was resolved by the migration? I can't see a preview so idk what it is

[neon epoch]

Yup :)

[drowsy zinc]

I hope I didn't close anyone's issues that were still pending

MM's sidebar issue had a thing about splitting off Administration/Development/... but I don't really want to do that, since you can just collapse those sections and not look at them

[vocal halo]

Yeah

[lilac edge]

@pliant canopy https://discord.com/channels/289587909051416579/555462289851940864 works on the discord client for discoverable servers, but doesn't on the web version

[pliant canopy]

I tested that on the web version lol

That screenshot I posted is the web one

[lilac edge]

that's weird... cause it didn't for me

using an account that's not in this discord server

[pliant canopy]

same for me

Well, I will add a link directly to the server. If it works for people they can directly join, otherwise they have the invite link as a fallback

[hard quartz]

well author and commit url not looks bad.

Subject: [PATCH] include commit url and refactor author
---
Index: src/components/overrides/LastUpdated.astro
IDEA additional info:
Subsystem: com.intellij.openapi.diff.impl.patch.CharsetEP
<+>UTF-8
===================================================================
diff --git a/src/components/overrides/LastUpdated.astro b/src/components/overrides/LastUpdated.astro
--- a/src/components/overrides/LastUpdated.astro    (revision 78b2191e4d303f89d79504a78698d78b3c1d1cde)
+++ b/src/components/overrides/LastUpdated.astro    (date 1745195088941)
@@ -1,10 +1,10 @@
 ---
-import { type GitHubAccount, getGitHubAccountFromFile } from "/src/utils/git-utils.ts";
+import { type GitCommit, getGitCommit } from "/src/utils/git-utils.ts";
 
 const { lang, lastUpdated } = Astro.locals.starlightRoute;
 const filePath = Astro.locals.starlightRoute.entry.filePath;
 
-const acc: GitHubAccount = await getGitHubAccountFromFile(filePath);
+const gitc: GitCommit = await getGitCommit(filePath);
 ---
 
 {
@@ -18,7 +18,9 @@
         })}
       </time>
       {" by "}
-      <a href={acc.accountLink}>{acc.displayName}</a>
+      <a href={gitc.author.accountLink}>{gitc.author.displayName}</a>
+      {" in "}
+      <a href={gitc.url} title={gitc.message}>{gitc.hash}</a>
     </p>
   )
 }
Index: src/utils/git-utils.ts
IDEA additional info:
Subsystem: com.intellij.openapi.diff.impl.patch.CharsetEP
<+>UTF-8
===================================================================
diff --git a/src/utils/git-utils.ts b/src/utils/git-utils.ts
--- a/src/utils/git-utils.ts    (revision 78b2191e4d303f89d79504a78698d78b3c1d1cde)
+++ b/src/utils/git-utils.ts    (date 1745195034362)
@@ -6,6 +6,14 @@
   accountLink?: string;
 }
 
+export interface GitCommit {
+  hash: string;
+  message: string;
+  url?: string;
+  author?: GitHubAccount;
+  date: Date;
+}
+
 const token: string | null = import.meta.env.GITHUB_TOKEN;
 
 const headers: RequestInit =
@@ -28,6 +36,26 @@
 const emailCache: Map<string, GitHubAccount> = new Map();
 
 // Git
+export async function getGitCommit(filePath: string): Promise<GitCommit> {
+  const hash = execSync(`git log -1 --pretty="format:%h" -- "${filePath}"`).toString();
+  const message = execSync(`git log -1 --pretty="format:%s" -- "${filePath}"`).toString();
+  const date = new Date(execSync(`git log -1 --pretty="format:%at" -- "${filePath}"`).toString());
+
+  let gitCommit : GitCommit = {
+    hash: hash,
+    message: message,
+    date: date,
+    url: "https://github.com/".concat(repo.concat("/commit/", hash)),
+  }
+
+  let gitHubAccount = await getGitHubAccountFromFile(filePath);
+  if (gitHubAccount != null) {
+    gitCommit.author = gitHubAccount;
+  }
+
+  return gitCommit;
+}
+
 export async function getGitHubAccountFromFile(filePath: string): Promise<GitHubAccount | null> {
   const email = execSync(`git log -1 --pretty="format:%ae" -- "${filePath}"`).toString();
   const cached = emailCache.get(email);

was a little test for replicate the commit url and include the git message in title or url.. (not sure how can look for very large commits)

[opal flare]

is that going to work in production? with git commands?

[hard quartz]

in teory, i check a little docusaurus and is like the same (they run git commands)
and in previews i dont see issues (well preview for the original PR not my tests)

https://github.com/facebook/docusaurus/blob/67924ca9795c4cd0399c752b4345f515bcedcaf6/packages/docusaurus-utils/src/gitUtils.ts
but i think they make a one run command with all the info and then manage fields with regex.. unless i bad read that

[hard quartz]

https://docs.papermc.io/paper/dev/pdc/#itemstack
maybe can be good show first the new way (not using ItemMeta) and in final mention the historical way?

[vocal halo]

Is the new API experimental?

[mighty dirge]

afaik not
https://jd.papermc.io/paper/1.21.5/org/bukkit/inventory/ItemStack.html#editPersistentDataContainer(java.util.function.Consumer)

[vocal halo]

ah okay i was referring to something else

[drowsy zinc]

made it do something like that

though I'm not sure about the looks, I accept any suggestions about that

[pliant canopy]

@drowsy zinc I think it doesn't look good

It just kind of ends up like this if the name is not that long and looks weird

[drowsy zinc]

yes, I know that, that is why I am asking for suggestions

[pliant canopy]

I think the way @hard quartz suggested it initially is fine

I would push a commit which does that, but I am too stupid to wrap this text here

I pushed the change up, feel free to modify it, as the current way is a big ugly. Using an empty tag and all. But I am not versed well enough in the arts of astro components that I could make that in a clean way

[neon epoch]

Can we remove the purple decoration to mark it as visited? Looks a bit janky to me

[still apex]

this looks too long imo

[pliant canopy]

Sure, we can do that

[hard quartz]

Was for avoid make a PR for a PR xd
(Just a little Co-Author if add any n.n)

[pliant canopy]

@drowsy zinc when you end up merging my thing, pls add @hard quartz as a co author thank you

[hard quartz]

Maybe this is better that a whole line (?

[pliant canopy]

I very much dislike it tbh

Or at least make it be directed to the right and be the same size text

[hard quartz]

The git info are more little?

[pliant canopy]

Yeah that's a smaller font size it seems

[hard quartz]

Directed to right?

[pliant canopy]

Aligned to the right

[drowsy zinc]

for any paper people in chat, I need someone to take a look at docs#574 (mainly at the configuration descriptions)

[brazen moon]

can you add the missing rotate vanilla permission?

the roadmap need to be updated too api stack only are not a thing anymore

[neon epoch]

I did give the descriptions a read, grammatically all seemed good.

[drowsy zinc]

which one is that?

[brazen moon]

on the permissions page: https://minecraft.wiki/w/Commands/rotate

wrapping is weird here btw

[hard quartz]

i dont notice that

[drowsy zinc]

the roadmap talks about making it an interface though, currently it's still a class but just delegates to CraftItemStack

so I guess I just change the wording to mention that?

[brazen moon]

yes the interface part is still relevant

[hard quartz]

better 🙂

[pliant canopy]

Btw scorp has officially overtrown @neon epoch from being #1 on the docs. Congrats ig

[brazen moon]

i think this section should really just say that at some point the class will be migrated to an interface, but there's already no api implementation anymore, they are all backed by an internal object. The constructor just call ItemStack#of

[hard quartz]

103k of deletes... wow

[pliant canopy]

(Only because he bumped dependencies )

[still apex]

can this be changed? https://i.imgur.com/WFJlrzs.png

being on the development page and the administration dropdownw being open/shownw

[drowsy zinc]

you can collapse it

[still apex]

thats annoying to do

[drowsy zinc]

the sidebar and pages are independent so no it can't really be changed

if you have a topic page open it will scroll to it on load though

[vocal halo]

I don’t know how far you are and if it’s still relevant but I also like this more. As some of you said, it might look better by aligning it to the right.

[pliant canopy]

It's already merged

Ended up looking like this in the end

[rigid bluff]

Is this docs: https://papermc.io/downloads/all ?

[spice temple]

Mmh?

[rigid bluff]

Is this page docs related ?
Because I have some improvement suggestion

[drowsy zinc]

it's the website

https://github.com/PaperMC/website

[spice temple]

But this channel is for all misc website stuff too, yes

[rigid bluff]

This is currently the case when you go to the Build Explorer for velocity or other paper products. You always end up at paper instead of the actual products

[still apex]

wdym

[rigid bluff]

https://papermc.io/downloads/velocity
Click here on build explorer

[eager plover]

it just takes you to the explorer

there is no automagical way to take you down to velocity when going to that page, and it's just really not a priority

[spice temple]

Would be a really simple pr tho to add a query param there 🙂

[opal flare]

@drowsy zinc maybe under "you may” add "use the links provided directly without rehosting” - some people think using hosted images without permission isn’t okay, in this case it’s fine

(But in better English)

[drowsy zinc]

I'll put it in the overview paragraph since it applies to all images there I presume

[opal flare]

yeah

[spice temple]

Idk who manages this but pls fix, new docs broke search?

Or do we not use it anymore?

[drowsy zinc]

only pencil uses it

[spice temple]

How does search on the new page work?

[timid cedar]

ik we're in the process of merging docs, but we've now good very good adventure localisation docs we could link to https://docs.papermc.io/paper/dev/component-api/i18n/ https://docs.advntr.dev/localization.html

[hard quartz]

Dont use that crawler stats? (Ex: the week email for most search things)

ref if still need https://www.algolia.com/blog/engineering/algolia-docsearch-astro-starlight-part-2

[drowsy zinc]

pagefind

[hard quartz]

https://pagefind.app/
Based in docs, for static sites

[drowsy zinc]

idk what you mean

judging by that email it should be possible to just proceed in the algolia admin ui without any changes

I don't have access to it so I can't do it

[hard quartz]

Yeah maybe just need in Algolia run a reset or something but the thing is i remember when comes from docusaurus you cannot touch things in the admin site and just let run the plugin

[pliant canopy]

Noted 👍. Might as well directly add that to the merge PR

[hard quartz]

well the basic docsearch works fine but in dev the search just points to prod site xd

[neon epoch]

Lemme have a gander

It only generated 134 records

Search will be neigh useless

[hard quartz]

then is valid try to add again the docsearch or just deprecate pencil search because i think the pagefind not allow search from external things (?

[neon epoch]

I’m gonna have a look

If we can get the crawler to work then we can keep pencil

[drowsy zinc]

https://docsearch.algolia.com/docs/templates#astro-starlight-template

you should probably use this crawler template

[hard quartz]

i just use the plugin and works

[hard quartz]

only issue is the search points to production

[drowsy zinc]

the site plugin is only a client for the algolia api

it doesn't have an effect on the indexability of the site

[hard quartz]

ahhh ok now i remember what you mean

[neon epoch]

yeah looking at the index it is basically just 1 item per page

[hard quartz]

its not like the same than docusaurus or not?

[neon epoch]

no i need to change the config. Now turning on my laptop to do so

[neon epoch]

yeah the default is worse than the docusaurus one lol

ok it picked up 1800 rather than 11,236 records now

I imagine that is because it is not crawling the config docs

[spice temple]

Thanks olli

[neon epoch]

nws should be done

It has significantly less records but all the config stuff is included

sorry for the email spam

[hard quartz]

email for any failed crawl attemp? xd

[odd harbor]

https://docs.papermc.io/paper/vanilla/#_top

[neon epoch]

a couple... yeah lol

[odd harbor]

https://docs.papermc.io/paper/dev/command-api/basics/argument-suggestions/#example-suggesting-amounts-in-a-give-item-command

[pliant canopy]

Yeah seems to work fine

[odd harbor]

https://docs.papermc.io/velocity/dev/plugin-messaging/#_top

[hard quartz]

well i supose is intended

[neon epoch]

This #_top is annoying

But it’s what is assigned to the title I think

[hard quartz]

yeah but unless the crawler can remove that when process the other way its just make Pencil "cleanup" that things

@neon epochthe crawls template used just modify the actions not?
because the docs not has a sitemap?

[neon epoch]

the docs do have a sitemap

and i have been modifying the actions

[hard quartz]

https://docs.papermc.io/sitemap-0.xml ok i found xd

[neon epoch]

it's dumb because there are loads of different records that link to the page and it chooses the one with that dumb #_top

[still apex]

can’t you just change the bot to not suggest that

[hard quartz]

that is easy but looks ugly if the issue is how the crawler is what has that url

[hard quartz]

that URLs are in the path explorer for the cralw too not?
ex..
the last time i touch the algolia was in a migration xd

[neon epoch]

nah

they dont show like that

[neon epoch]

right back at this now

im assuming i can probably just delete it with the cheerio object

im gonna crash out

[hard quartz]

._.

[neon epoch]

do i just remove these one by one until it works

[hard quartz]

one by one of what? xd

[neon epoch]

$('[id="_top"]').removeAttr("id");

and then remove, starlight__on-this-page--mobile

etc

[hard quartz]

xd

[drowsy zinc]

is that with the starlight template?

[neon epoch]

yeah

[drowsy zinc]

weird

[neon epoch]

with some modifications so that the config blocks index

[hard quartz]

well also pencil need a update for the DocSearch library... https://www.algolia.com/doc/libraries/java/v4/upgrade/

[neon epoch]

Basically they have the "_top" id there. If i remove that it just populates it with the starlight junk

that stuff is there on other elements either way

[hard quartz]

well the main docs of starlight has the same thing but well they not use Algolia

[neon epoch]

lol... surely this is not intended behaviour

[neon epoch]

Try the bot now

[odd harbor]

https://docs.papermc.io/velocity/dev/plugin-messaging/#case-2-sending-a-plugin-message-to-a-backend-server

https://docs.papermc.io/paper/vanilla/

https://docs.papermc.io/paper/reference/world-configuration/#anticheat_anti_xray_use_permission

[hard quartz]

i dont see that now

[neon epoch]

looks to be working

[hard quartz]

just with the record-extractor for the crawl not?

[neon epoch]

yeah just removing those couple of starlight id's fixes it

[drowsy zinc]

it shouldn't pick up those at all though

[hard quartz]

oh i was think you just remove the href xd

[drowsy zinc]

they're outside the CSS selectors

[still apex]

is @odd harbor broken

https://i.imgur.com/RHhFgKZ.png

[neon epoch]

It could be my fault

Or it could be pencils fault

[grand dew]

the docs stuff doc knows about and is fixing at somepoint

or did fix afaik

[hard quartz]

Me?

[grand dew]

yeah

[hard quartz]

@grand dew i think its just freeze or offline because the google command fails to

[grand dew]

ah

[young matrix]

Someone needs to move it's functionality elsewhere or determine why it started stalling rather than just restarting it every once in a while

But that person is unfortunately not going to be me, at least for the foreseeable future

[still apex]

(what i've been saying for a while :p)

[spice temple]

I refuse to touch discord shit

But ye, the restart thing is obviously just a bandaid, but it works good enough

[neon epoch]

Every discord bot I’ve written kills itself at intervals

Idek know why

I just get random exceptions from stuff normally

[eager plover]

Discords endpoints are stupidly stable

A lot of the libraries generally don’t have a good means of handling that

[hard quartz]

how handle the gateway always is a issue.... for the reconnection thing also

[vocal halo]

Forgot a subscribe somewhere

Is the code OSS? What language is it written in?

[still apex]

pencil is open source

[neon epoch]

https://www.github.com/e-im/Pencil

[still apex]

www

[neon epoch]

Had to slowly fix that link

[still apex]

@odd harbor

still dead

[spice temple]

Not still, that's not how this works

It's dead again

It will undead itself in a bit

[hard quartz]

XD maybe can try PR an update in pencil for last SNAPSHOT with the update in Reactor core/netty and see if that helps if the issue is maybe how netty handle that dead things

[spice temple]

Sure

[hard quartz]

If remember pencil only need handle one thing in the update for the reference message

[spice temple]

Idk if it still auto deploy, ping me once it lands so I can check

[hard quartz]

Okay, i dont perms in pencil but i make the PR for that

[hard quartz]

ready https://github.com/e-im/pencil/pull/4
i cannot make run (like in docsearch) but the only breaking thing was the message reference for attachments

[lilac edge]

is that really how it needs to be done now? that's so ugly

this is pencil so it doesn't matter that much but bleugh

[hard quartz]

Mostly of that are optional but the old way was just an Id and just confuse with how discord wanna things
Maybe can create a spec in D4J and allow pass directly a message or use reply

[twilit fog]

Shouldn't entities be listed here?

cause an actual server has them like this

this also differs from the docs

is this correct not saying "to 1.21" in bukkit.yml docs?

[vocal halo]

What would you replace it with?

[twilit fog]

which one?

[vocal halo]

The last one

[twilit fog]

"from 1.13 to 1.21"

[brazen moon]

yes and the minor part looks outdated, api-version now support that digit, should be dynamic too or just 1.13+

[twilit fog]

might give the wrong impression as in someone might write 1.20.4

[brazen moon]

it should just say to follow the api-version format or smth, feel free to open a PR!

[brazen moon]

for this one maybe mimic the despawn-time section

[vocal halo]

Just to make sure, we are talking about bukkit.yml and not plugin.yml is the behaviour identical here?

[brazen moon]

yes it's the same

[vocal halo]

👍🏻

[random pollen]

@jagged pecan ok wait, the intellij errors I saw before aren't there but check isn't actually fixed https://pastes.dev/BGkMYJGAtL

I think thats the same thing I saw in ij, not sure why it didn't show up in check before

[jagged pecan]

Huh, interesting, it satisfied vscode + tsc

[random pollen]

maybe it short circuited on those other 3

[jagged pecan]

If I get time, I'll try and see what I intellij does differently

Was that idea + the relevant plugins? Or smth like webstorm

[random pollen]

this isn't intellij

this is bun run check

after applying your diff

[jagged pecan]

Huh what

[random pollen]

it removed the other 3 errors but adds this one

intellij was just showing it before and after the diff

[jagged pecan]

for some reason github gave me 0 notifications around that pr lol

[random pollen]

https://github.com/sveltejs/svelte/issues/15182 ok why is this guy having the same problem also building a tree component

[jagged pecan]

huh

rebasing to master, this patch still seems to produce no errors for me, with both bun check tsc ., or as an error line in intellij

so i have no idea whats happening

Index: web/src/lib/components/tree/Tree.svelte
IDEA additional info:
Subsystem: com.intellij.openapi.diff.impl.patch.CharsetEP
<+>UTF-8
===================================================================
diff --git a/web/src/lib/components/tree/Tree.svelte b/web/src/lib/components/tree/Tree.svelte
--- a/web/src/lib/components/tree/Tree.svelte    (revision c55247b1bd38a869a50245ecd1f12273154292e9)
+++ b/web/src/lib/components/tree/Tree.svelte    (revision e375ea2642ff887d43d572c5ca01eca969f77d7f)
@@ -1,5 +1,5 @@
 <script lang="ts" generics="T">
-    import { type TreeProps, TreeState } from "./index.svelte";
+    import { type TreeNodeView, type TreeProps, TreeState } from "./index.svelte";
 
     let { instance = $bindable(undefined), roots, nodeRenderer, childWrapper = null, filter = null }: TreeProps<T> = $props();
 
@@ -16,7 +16,7 @@
     }
 </script>
 
-{#snippet renderNode({ node })}
+{#snippet renderNode({ node }: { node: TreeNodeView<T> })}
     {@const collapsed = requireInstance().collapsedNodes.has(node.backingNode)}
     {@render nodeRenderer({ node, collapsed, toggleCollapse: () => requireInstance().toggleCollapse(node.backingNode) })}
     {#if childWrapper !== null}
@@ -29,7 +29,7 @@
     {/if}
 {/snippet}
 
-{#snippet renderChildren({ node })}
+{#snippet renderChildren({ node }: { node: TreeNodeView<T> })}
     {#each node.visibleChildren as childNode (childNode)}
         {@render renderNode({ node: childNode })}
     {/each}

[random pollen]

look at master, I just pushed that exact diff

[jagged pecan]

ah ok

[random pollen]

maybe different import order idr

[jagged pecan]

yeah im slow?

[random pollen]

huh?

[jagged pecan]

import order shouldnt matter afaik

[jagged pecan]

i pulled master before re-appling, but sometimes i just take a while fighting git :)

[random pollen]

yeah I don't think it will here

I just stashed all my local changes and did bun check with latest master, still getting the error

if you're not getting it maybe it's system specific somehow

I'm on linux rn

[jagged pecan]

same. what node version is intellij configured with? mine is node22 via nvm

oh wait

[random pollen]

I'm just running bun check from command line, I'm semi accustomed to ignoring errors in svelte files from ij since it doesn't work properly with ts in templates

❯ node --version
v23.9.0

❯ bun --version
1.2.11

[jagged pecan]

yeah bun check has nothing for me, my bun version is 1.2.9

[random pollen]

let me try nuking node_modules

[jagged pecan]

ill try the self-upgrade now

[random pollen]

ok nope I nuked node_modules and reran install, same thing

[jagged pecan]

ah i nuked node modules, reinstalled, and now i get the error

interesting

[random pollen]

I commented on that issue, might be worth mentioning that there, idk

https://github.com/sveltejs/svelte/issues/15182#issuecomment-2848325692

[jagged pecan]

let me re-downgrade bun to make sure that wasnt the issue

yeah no not caused by the upgrade, but by the node_modules re-install

[random pollen]

and you were already up to date with bun install before nuking it?

should def mention that on the issue since it seems like the maintainers were having a hard time reproducing it

[jagged pecan]

probably not, thinking about it. let me test by re-installing against a previous repo state

yeah my dependencies are quite old comparatively

I had installed just before that warning started appearing, in commit 77bc92

according to bun, which was instructed to install from the frozen lockfile, the following packages updated

+ @eslint/compat@1.2.7 -> 1.2.8
+ @eslint/js@9.22.0 -> 9.24.0
+ @sveltejs/kit@2.19.0 -> 2.20.5
+ @tailwindcss/vite@4.0.13 -> 4.1.3
+ @types/diff@7.0.1 -> 7.0.2
+ @types/luxon@3.4.2 -> 3.6.2
+ eslint@9.22.0 -> 9.24.0
+ eslint-plugin-svelte@3.1.0 -> 3.5.1
+ svelte@5.23.0 -> 5.25.10
+ tailwindcss@4.0.13 -> 4.1.3
+ typescript@5.8.2 -> 5.8.3
+ typescript-eslint@8.26.1 -> 8.29.1
+ vite@6.2.1 -> 6.2.5
+ @octokit/openapi-types@24.0.0 -> 24.2.0
+ bits-ui@1.3.13 -> 1.3.19
+ luxon@3.5.0 -> 3.6.1

specifically, upgrading svelte@5.23.0 to svelte@5.25.10 svelte@5.23.1 is the root cause
edit: narrowed it down even more

[random pollen]

interesting

[jagged pecan]

https://github.com/sveltejs/svelte/compare/svelte%405.23.0...svelte%405.23.1

whats interesting is that issue you linked, apparently https://github.com/sveltejs/svelte/pull/15473/files was opened to maybe fix it
which first shipped in .1
so it might have made it better for some, but this could well be what broke it for this?

[random pollen]

maybe, I'm not sure

would have to test

I think the OP's diagnosis was a bit of a red herring (or rather acted as one for the svelte team)

becuase they are right, no one will do that weird import thing

this happens just writing normal svelte

[lean venture]

why is the talk about svelte?

lol

thought docs were in astro

[spice temple]

this channel is more than just docs

its what we use for all misc web projects, like patch roulette in this case

[lean venture]

OH okay

[pliant canopy]

Astro also uses svelte

Or, well, at least the docs do

Svelte ^^: https://github.com/PaperMC/docs/blob/main/src/components/tools/ItemCommandConverter.svelte

[spice temple]

astro has 0 frontend js

you can use whatever you want with astro

but you are not gonna get good integration between astro components and framework components

which why I will never understand astro

[neon atlas]

@drowsy zinc https://github.com/PaperMC/docs/pull/589/commits/c4df9a0be8cb6e9a632e6c3893e7c0affe0a2ca1

[drowsy zinc]

vanilla big v otherwise great

[neon atlas]

[vocal halo]

what is this embed?

[south thunder]

mja's profile pic

[pliant canopy]

somebody who is very happy about the current state of the paper documentation

speaking of which, I should finish my PRs

[vocal halo]

im more talking about why it looks like this. usually if you link something on github it has that standard embed

with name and stats

[pliant canopy]

It has been showing the profile picture on prs for a while now

[still apex]

https://github.com/PaperMC/docs/pull/589/

only on commits?

[vocal halo]

maybe

[neon epoch]

I feel like that’s intended behaviour can’t remember tho

[vocal halo]

well it looks unfinished idk

[jagged pecan]

i mean what sorta integration did you want? it seemed pretty seamless for me, including ide integration etc when i did an astro site with a vue client sided component

[eager plover]

I think he means that astro and framework will be using different components and will likely involve some headache of either manually keeping stuff in sync between astro and your framework

[vocal halo]

Isn’t it like that if you are using a vue component you can’t use an Astro component inside of it or something like that? I remember having issues with that (I’m not an expert and did it merely to try out Astro maybe I’m just too unfamiliar)

[eager plover]

that would generally be what I'd expect given that vue and other frameworks generally like to understand what is going on and not have random things pulling in html/wahtever from elsewhere

[spice temple]

No Vue is fine with that

I do that at work