#docs-website
1 messages · Page 2 of 1
yeah i got rid of it
If it's circular loading the stuff should be in a circle.
Also, could someone with commit access look at #182 and #181
They both look good to me
Please replace blocks and arrows with spider-man pointing meme
Could actually be a good idea. Everyone understand that meme
how exactly does this work? Are these dependency things like checking if PAPI is in the server or what exactly?!?!
the phrase "cyclic loading" makes me think "slow loading"
whereas when i hear "circular" i think DI
that seems like a you problem 
well even in the message it says circular haha
i mean, cyclic is the term used in most places because it refers to a cycle
which, applies if you use it in terms of graph theory
cyclical?
this is clearer i guess
when i hear cyclic id think about the gc or something
I don't know where you get that from but... I guess if it helps people? 😄
doesnt really matter honestly
you could throw any error and most people would be able to figure it out either way
this is all it takes to explain
yeah - i see what you mean though about it being inconsistent between cyclic and circular
I would rather stick to cyclic(al) though as this is the proper term so may change the circulars to reflect this
ye i dont exactly know academic graph theory
tbh many conventions in math tend to differ from the way we expect things to be
such as Z being treated as a vertical axis, i think the the way mathematicians think vocab through is just different
Folia Docs are currently mismatched between dev / admin. Would you like these pages in admin or development.
based on the slugs, they should have been placed under admin
What's the point of the version_group/family endpoints in bibliothek?
or, at least the intentions
1.20 is a group, 1.20 and 1.20.1 are in that group
@neon epoch what do you mean here? https://github.com/PaperMC/docs/pull/145/files#r1197573053
it seems to work fine?
ermmm, it’s odd that it works tbh. In MDX 2 you are supposed to need [] around the text so like:
:::note[custom text]
:::
I don’t know why it compiled but the rest of the codebase is in this other format now @echo canyon
ok, I will change all of them
I think its still based on a pre-mdx2 commit hence why that works
oh I didn't rebase you mean... yeah
should I add an index page for the Component API group?
didn't realize you could do that, but I see the API category does have one
that's the README.mdx file at dev/api/ right?
Yes
I don't think it matters too much at least for now, but you can add one if you want
man, I'm still not sold on how the "config reference" pages look. even if you make the sidebar include all the levels of settings, its just so cluttered
I think I'm imagining something that looks like the exact yml file, but you can like click on settings to expand extra info about them
but ctrl+f will search through all the text still or smth
like isn't it super difficult to pinpoint where exactly the "secret" setting is in the hierarchy?
you have to look up and find the slightly bigger text, and then look on the right to see the category further up
and maybe its not visble so you have to scroll up, and bleh
I like your point about the YML thing, but I am pretty sure that's not going to be possible / easy. Might be possible to add some sort of like hover and it gives you like the root.node.leaf yaml format
What we really need is somebody to go comment the configuration classes
and then we can write a tool to dump that stuff out and then we can fight over the site layout
that really isn't important for this, the comments for the yml are all already written, they are right there on the docs page
are all nodes commented
I don't know the ins and outs of docusaurus, but its not possible to completely dynamically generate a page based on a json file of all the settings + defaults + descriptions?
not at the moment, but that's just cause configurate-yaml doesn't support comments yet. The comment language can just be copied from the docs
I thought cat meant the java class
it would probs need some manual intervention, i.e. somebody to actualyl do that, or some task which just tries to form the page manually
I did mean the java class, nothing in there is commented atm
commenting in the java class is easy, its just copy/pasting from the docs page now.
i mean, you could move the comments over to paper and then have a GH action that scrapes them
I do want us to be able to eventually auto populate that thing in some degree, even if it's just a case of running a command once a few weeks
right now stuff generally tends to fall out of sync between the two
yeah, but that still doesn't really have anything to do with what I said, I'm not talking about the syncing or whatever.
yeah, but that's pretty trivial
ik, but I think it would give us the means to re-do the page layout
I'm inclined to agree that the current layout is far from optimal in terms of stuff
which is what I was originally talking about
was some level of "would be easier to try different things if we're not having to manually tweak it and determine how we feel"
idk
If someone draws up an idea im happy to have a stab at it
See, part of me isn't sure if one can really make something nice with docasaurus with that
like, he issue here is that the docs page doesn't really make subsecions stand out as such, everything just kinda blends together unless you're aware of the layout already to a degree
it's not really designed for crazy levels of nesting
can't you just make a jsx component and import it?
if you can do that, it will be really easy
just completely ignore "markdown" and make a react component that functions like we want
Yea, I think so, but idk exactly how you'd represent "hey, this stuff is at level 1 nesting and this stuff is nested 200 levels"
like this maybe
my idea was for it to look exactly like a yaml file
like ides do
yeah I'd just do it as a react component
my idea was for it to look exactly like a yaml file
and then, when you click on a setting, it expands a dropdown below and has the full path, and the description
but ctrl+f has to still work to search the descriptions somehow, not sure how that works but it feels possible
if the text is just hidden, does it show up with ctrl+f?
I think not
depends on the browser and how its hidden
I've had some weird experences on that front
well we could just take over ctrl+f for that page, or add our own search bar
hmm seems like it does
but yea, I think it's doable, idk if you can get the selection and expand the tree easily, I'd imagine you can?
at least on chrome, with just <details>
if that works on chrome, firefox, edge and safari where it automatically opens it, we can absolutely use that for hiding the descriptions
doesn't work in safari
not firefox either
seems like its just a chrome thing
and I really don't think taking over ctrl+f is a good idea. personally I hate sites that do that
we could add an extra search box but meh
well we could literally just do a yml file with comments then
I feel like that is already significantly better than what exists now
yeah, it is pretty bad
would be nice to at least be able to anchor link to each element though
yeah, so we couldn't do that just with markdown syntax highlighting then
If you do display none I don't think browsers show it in control f
If you change the opacity it shows up
So you could do it, idk if you can detect if stuff is highlighted via Control f tho
Yeah no you can't
how often does pencil re-cache or whatever the list of docs pages?
tried to search for the new component page and headers and didn't get anything
oddly specific time
what is the audio in that 🤣
but yeah, that's the idea
you can put the default as the "value" for each key too, and if its gonna look like yaml, then it can have the indented structure instead of the . separator
oh lmao I didn't realise it captured my other tabs LOL
@echo canyon
@neon atlas (im pinging you cause I see you are online) What do you think about where we should put the description. It currently just gets tagged on the end but im not sure if it should almost be like a dropdown inline with the yaml?
Yea I can see how info at the bottom is going to get out of hand for larger files
👍 definitely
Maybe a click can expand a yaml formatted comment ontop?
not sure i follow
Also, stuff should work with anchors so that you can still copy urls that directly point to one setting
hmm, that is gonna be a pain but will look into that
done this
Also need to think about SEO, the text always needs to be in the Dom so people can search for it using Google
i mean, could that be as simple as just having a complete transcript hidden somewhere
I'll probably need some help with the anchors and SEO - but that can happen when its in a PR or pushed.
Don't make me touch react again :/
mate i have literally learnt react this morning
to be honest - the thing i’m not looking forward to is extracting all the existing docs
Oke, this is a huge WIP but yeah... see PR comment
so the whole "searching" thing. I think its important to have some way of searching. If we can't find a way to integrate ctrl+f to somehow automatically open up descriptions, then we probably need just a separate simple search bar?
I like the look of it so far, tho. Couple things tho. I think it should be more obvious that you can click on settings to expand a dropdown. Maybe even an outline of bg color change. Also the dropdown should have a separate background to disinguish it from the actual settings
I also don't really like descriptions being hidden until you click them
More often than not you need to look at multiple things at the same time, having to open them individually makes it tedious. meh idk. but the test looks considerably worse than what we have atm to me thanks to dropping all formatting within the yaml-like structures.
smaller thing but you cant copy description text as clicking closes it
the latter
you think it looks worse that this? You can barely tell where these settings are. If you are just talking about changing the formatting of the description so it differs from the actual yaml, then I agree and already said as much.
If you can add extra margins, and a similar bold/code formatting then it’d look good yeah
What do you think about the “search” aspect? Apparently only on chrome can you design smth that auto opens when you use browser ctrl f
Not sure what the best solution to what would be
as sulu said, yoinking ctrl+f behavior sucks, but I think adding a simple search bar would be good enough, although kinda weird
Or just have everything expanded by default with a collapse button
that might be nicer
what if we listened for ctrl+f and expanded everything then?
so not hijacking it, but making it work if someone tries to do it?
along with collapse/expand buttons
All Hail Svelte
Also, imo this is quite a visual clutter. Think it should be highlighted the on currently selected, only one can be selected at a time and I don't think it should be inside the yaml block, it should either be display outside like before or on hover imo.
I think that all is solved by just making the desciption box not have the same background. give it a different bg, a border, bold property names, and boom
maybe...
i mean, i had some time so i just threw something together. i’m a complete noob at this. I’m not sure if we are able to easily format once it is inside a YAML code block - will have to look tomorrow
if anyone wants to PR to my branch feel free :)
i think i agree with machine more vs kenny though- the old design - although very simple - lacks all ability to show easy inheritance. Also, there are a fair few nodes that lack docs.
Also not sure with what to do with the source yaml. Currently is just plonked in the mdx file but, 
Supports markdown now
Searching is done - it only searches the titles currently though
Should prolly grab some fuzzy search algo from somewhere
I threw together the worlds worst python script to convert the markdown to yaml; https://paste.ollieee.xyz/exexesapuy.py
Here are the deployments
world Config:
https://docs-dyrla5goq-papermc.vercel.app/paper/admin/reference/configuration/new-world-config
global config:
https://docs-dyrla5goq-papermc.vercel.app/paper/admin/reference/configuration/new-global-config
Search needs some work but I think its more readable
https://i.imgur.com/MG93uon.png imo this could look a bit better, when you click on a config option
i don’t disagree, it’s just coming up with something
still dont think you should be able to preview more than one configuration
might be confusing, idk
I think the “default” shouldn’t really be a thing there. Can’t it just stay where the value usually is? So only the description is hidden?
could be in () behind the key?
oh what do keep the default inline with the node and only drop down the description
err, no. maybe making it more obvious which are expanded though
up to you, I personally prefer only having one thing opened at the time
@jovial shell how do you feel about continuing your docs PR for plugin messaging or would you prefer someone else to take it over?
Anchors are now implemented logic wise, just need a button to copy them. More style changes and also search removed cause it was slow and laggy. Search is the main thing left at this point
with the anchors, it'd be nice if a config option was opened, if it had a button to copy a link to that anchor
also, the indents look to be different sizes?
I'm wonder what "paperweight userdev" should be called in the docs? Should it be:
- Paperweight Userdev
- paperweight userdev
- Userdev
- userdev
paperWeight User_DEV 💅
Tbh I like the first one tho
they are slightly i think, will investigate
was planning that it would be like the normal anchors - so maybe when you have enabled the drop-down, it will show a # that can be clicked to copy the path to clipboard
sorry
not the right channel
where?
#paper-help if you are using paper
thx
Why are these files in my changes list?
add .idea to your .gitignore
how?
Doesn't intellij generate a gitignore in the idea folder?
only if you have the plugin I believe?
Oh that could be it
this is a little confusing, ambiguous
you can't just rename your plugin.yml to paper-plugin.yml and be good to go which is one interpretation of this I think
true
yeah
Hi, I had some time off from work this morning so I have updated my Docs PR's
my config stuff is at the stage where I am happy for any proper reviews and public feedback now:
https://docs-ty89r7zci-papermc.vercel.app/paper/admin/reference/configuration/new-world-config
Have addressed the comments on #183 (paper Plugins)
Have addressed the comments on #187 (debugging and stacktraces). Just not sure about MM's comment about where the reading stacktraces should go - Dev / Admin
#185 Needs some input, Idk what is needed other than the Schedulers?
bump @jovial shell :)
I guess its fine to be in dev, whatever
could symlink it to admin too, specifically the reading stacktraces one
but more importantly, I think that the slug should have an extra path part
like paper/dev/smth/reading-stacktaces. Not sure its good to have all of them at the root dev/ right?
althought, I'm not seeing, that all the pages that weren't added by me are just at the root dev/... so idk
This is kinda why i just wanted a wiki, as wikis know how to handle this stuff already
I don't have hardware to look at stuff, but, is there a note about the omit stacktraces thingy?
no, what’s that?
There's a JVM flag, omit stack trace in fast throws or something
The JVM will basically not generate a stack trace for some exception throws when the JVM optimises code, that's why you get stuff like npes without a stack trace
Yea
Was thinking it may be worth making a note of that somewhere given how often it's tripped people up
re: plugin messaging pr, should BungeeCord be used or bungeecord:main? since both of them are used in the code examples
I would not main bungeecord:main, bit of namespacing is always good
is there actually a difference? imo if there’s no difference then it should stick to 1 and not multiple
er
so
the problem here is that spigot rewrites the messaging, and so you can't use bungeecord:main from inside a spigot plugin
idk why tf he didn't at least make it directional
so… Capital BungeeCord everywhere?
Well
here is the fun one
for spigot and co, use BungeeCord
outside of that, you need to use bungeecord:main for 1.13+
someone has ready files optimized for 1.20?
Community contributed configuration guide for Paper: https://paper-chan.moe/paper-optimization
Take a look at this for suggested configurations
i already have this but i mean .yml ready to put in
I'm not aware of those
Every server is different. The goal is to find one works for your use case. Blindly copying something is bad.
@neon epoch did we wanna fully switch to "BungeeCord"
oh yeah forgot about that, im just reviewing nacios PR and then will take a look
I would use that channel and just note that outside of paper, that can change?
Idk the context of the doc entirely, so
Plugin Messaging guide

Done 🥰
I'll have a look in a bit, should then hopefully be good to merge 🙂
cool cool 
smacked some last review points, if you ping when when you are done I'll smack the cute green button
yeye cool cool
watcha talking about here @neon atlas
er, well
the channel name for 1.13+ is bungeecord:main
but, spigot has some janky ass rewriting of that, so you need to use BungeeCord inside of plugins ontop of anything based on spigot
yeah ok, i just meant generally why attach the comment to this line
I mean, i'd imagine he just clicked somewhere
👍
idk, I generally don't expect much from lynx ||jkilybb||
I did just click somewhere
cat knows me too well 
I had a though (those are rare) and before it left my brain I just smacked it onto the PR
You had a though?
fuck off, its warm today my brain doesn't do much
why do you think I am reviewing docs PRs 
Lmao
Well no the channel originally was called BungeeCord no ?
and then got moved to bungeecord:main
yes
the server still expects BungeeCord tho from plugins
because 1.13+ changed it to keys
and then fixes that for plugins
why do we even mention the bungee channel 
forgot the method name and not even gonna try on this craptop, but theres a method which changes the name from BungeeCord to bungeecord:main and vice versa, so, if you try to use bungeecord:main to send on spigot/paper, it blows up
cause it expects a key style channel
well, because plugins .equals for BungeeCord
idk why he didn't make the method directional so that you can at least send properly; idk, replacing that API is on the list
The velocity page uses bungeecord:main throughout
Yea, it's, weird...
i think im happy with that though, its right at the top so its explained before you send anything
so yeah, @neon atlas done
noice 🙂
wow, really not gonna approve it cause you are playing spigot putt party 
never
There ya go
never even reviewed my PRs 😔
what version should I use?
no idea
LTS works
Could someone check the build log of https://github.com/PaperMC/website/pull/53 to see why it's failing?
nvm it just needs to be approved
Yes, just did that
thanks (:
ok wait, I clicked approve too but it didn't do anything
if I didn't have perms, why could I even get to the approve button
I did first too, think I did the wrong link
It looks like it was approved in the past and just lost it or something
so i have a flight tomorrow, what docs need writing 
if you want to do adventure stuff, theres a bunch listed on an issue
have you paid any attention to owen's command stuff?
although idk how solid that stuff is yet, so idk if writing docs now is a good idea
Plugin Configuration files means like JavaPlugin#getConfig ?
could also look at a simple jdbc demo
i might just end up watching netflix, but just wondered if i should consider writing these?
was gonna say, jdbc would be a good one, especially around velocity
generally, in the long run, we want to replace all of the OG bukkit wiki
LOVE YOU
watcha mean about velocity? i was gonna just make a general “this is how to create a connection”
and maybe like something about DAO type objects and a wrapper class to handle them
🥰
i think that there was some classloader caveats
minimisation ofc breaks stuff
and then just a general "how to mysql" would maybe be nice, hikari go brr
Ok, so i will probably look at configs and databases assuming i dont just like fall asleep or get distracted by the office lol
Did they fix the way to remove the End portal frame, so i can build a sand duper? (I play on 1.20.1)
I try it with Red mushrooms
//
Can i allow to use rail dupers and remoing the frame at the end portal?
this is not a support channel
hangar-help is ok for my question??
Oh, ok. Thank you.
Could somebody help me set up velocity?
i added a comment about configurate to the config guide but i don’t think i will bother explaining how to use it and just point to their docs / homepage. syntax might change so no point in doubling up
hey I'm seeing the same issue this posted in this forum https://www.spigotmc.org/threads/1-19-playerinteractevent-not-called-when-entity-is-in-sight-client-bug.574671/ but the resolution looks like a hack. is there a better way to handle the case where player interact event is not called when 4-5 blocks away from a mob looking at it?
If you are a developer, you should use #paper-dev This channel is meant for the documentation site of PaperMC
Hey @young matrix, was wondering regarding the docs repo, how does docusaurus know to use docusaurus.config.ts when using the dev script instead of docusaurus.config.js, I would of imagined building took care of this as typescript is transpiled into javascript but it doesn't build at this point?
that's a great question. it's... a little hacky
I wouldn't have done it this way if I wasn't already using yarn patches to patch docusaurus for other things
so wouldn't really recommend doing that in your own repo. what you can do is just create a dummy docusaurus.config.js (even not in the root and specify --config for everything in your package.json, but that can be a bit meh) and transpile and re-export there.
Yeh I was looking at doing this in my own project, was looking at the --config param but didn't really fancy it and noticed paper was doing some funky magic and tried to find out how but looks like I might stick with the custom config 😅 . Thanks for the explanation.
yeah, it's not ideal
I think one of the docusaurus developers was talking about adding official support at some point but not sure what came of it. haven't been following docusaurus too closely for the last few months
That should be in coming in 3.0.0 but it’s been in alpha for a while iirc
nah paper uses 3.0.0 (before it was in alpha, not sure if its been updated to the alpha). doesn't seem to be added yet https://github.com/facebook/docusaurus/blob/main/packages/docusaurus/src/server/config.ts#L16-L21
Oh right
so @neon epoch I think the anchors don't take into account the height of the top toolbar. So like linking directly to the option makes it too far up.
https://github.com/PaperMC/docs/commit/e5322b2a7b54e9c2255f7818734f0a4bca4c597b average paper em cee developer committing straight to main with a mistake in it.
@burnt ice
you did default twice and didn’t list ZLIB in the list of possible values
the option above it also has default twice
@echo canyon Im gonna get rid of the URL Copied message as it will be a pain to get in the right place and now that I properly ypdate the hash in the browser it doesnt matter anyways. Will continue to copy it though
There wasn't one before anyways

thanks for the merge
Just gotta work out what to do for this
Cool, this is the most up-to-date deployment:
https://docs-o93g09yj3-papermc.vercel.app/paper/reference/global-configuration#block_updates_disable_tripwire_updates
I'll just spin it up locally 😅
fair enough :)
diff --git a/src/components/Config.tsx b/src/components/Config.tsx
index 9444871..6d26171 100644
--- a/src/components/Config.tsx
+++ b/src/components/Config.tsx
@@ -123,9 +123,9 @@ const renderYamlData = (
renderedNodes.push(
<div
key={key}
- style={{ paddingLeft: `${root ? 0 : INDENT_SIZE}px` }}
+ style={{ paddingLeft: `${root ? 0 : INDENT_SIZE}px`}}
>
- {key}:
+ <div className={`config-auxiliary-node`}>{key}:</div>
{renderYamlData(
value as YamlData,
parentKey ? parentKey + "_" + parseNameToHash(key) : parseNameToHash(key)
diff --git a/src/css/custom.css b/src/css/custom.css
index f20efca..edf0ae0 100644
--- a/src/css/custom.css
+++ b/src/css/custom.css
@@ -206,3 +206,11 @@ html[data-theme="dark"] {
text-decoration: underline;
cursor: pointer;
}
+
+[data-theme="dark"] .config-auxiliary-node {
+ filter: brightness(60%);
+}
+
+[data-theme="light"] .config-auxiliary-node {
+ filter: brightness(500%);
+}
lookin good
People lurking in this channel, can you give your opionions on:
https://docs-rioa6t1c5-papermc.vercel.app/paper/reference/global-configuration
Please and thank you 🥰
not very mobile friendly
what in-particular because i have been using it on mobile throughout and i think it is fine?
oh nvm you need to press on the thing twice for the description thought you had to press the # only
yeah you do on mobile
I still see this happening. I didn't put it in my review since I already pinged you here, so idk if that was included in "all of machines comments fixed"
it seems to only happen the first couple of times and then it works fine :/
that’s for me on my phone at least
@echo canyon can't seem to replicate it on my PC hence i concluded it was fixed
ok, so on safari here, it takes like 1 second and then scrolls to the correct place
ugh, i just tried firefox and it always fails there
why is this so stupid
Need to account for the navbar somehow
Web development 
🥴
Clicking the # results in a scroll to this y value:
But then loading the page results in this:
OH... It seems like firefox has some more advanced anchor logic as if I comment out the scroll:
Firefox still scrolls to that stupid middle of the element location
But chrome does not scroll at all
im glad we all decided to implement it differently
@echo canyon I can't find any docs online on how to stop this behaviour so im not sure if this is a cantfix type of thing
Ok, so Firefox intercepts the hash changes in the browser and has its own (flawed) scroll logic. I can somewhat fix that by putting the scroll in a setTimeout so that it scrolls to the correct place after the incorrect browser one, but there is no actual way to fix it properly.
This has no sideeffects on chrome as it just scrolls to the correct place 1ms later on
i guess i could try and break their logic by making the hash more proprietary
WHAT IN THE ACTUAL FUCKERY
Using the hashcode now
Firefox is cringe confirmed.
When you just submit a hash change to the searchbar it doesnt reload the page, so SOMEHOW it knows to scroll a little bit down from where it should be eventhough the hash in the bar has NOTHING TO DO WITH what the string reads
im at a loss
And this is repeatable when all code to do with scrolling is commented out. Like what the actual
i've spent too long on this today. I have pushed all I have worked on, but if anyone is unhappy with the behaviour of it now, meh i dont know what to do
webdev do be like that sometimes.
it looks fine to me. The smooth scroll on page load to an anchor is kinda weird, if that was instant scroll instead of smooth it'd be better, but its really fine as is. idk how much customization of the scroll you can do
Scheduler docs have been superceded by me with permission from nacio. Ready for final review / merge
does anyone know if wardens count towards the global mob cap?
They do count to the vanilla mobcap like every other mob.
Wrong channel, use #paper-help
Hah, lovely rename to stop the daily confusion
prove it. /s
😂
@spice temple obviously I am british, so you will probably find my PR's will contain some british spellings. Seems that one instance of Colour made it in
Just something I noticed, shouldml prolly have guidelines on that
everything should be in French
yeah, will correct the one you commented on and that Colour in the plugin configs PR tbh. its just one word in a diff. file

Addressed your comments though mini
Good
how dare you destroy my heritage

@echo canyon i have thought it was Docosaurus this whole time LMAO
🤣
it's not?
WHY
that’s what i’m saying
Can i get some reviews on https://github.com/PaperMC/docs/pull/191 please
#203 is ready to merge and I can then update #183 once that is done
#204 looks like it can be merged as well, just needs an approval
@neon epoch some of your diff seems to be removing the #anti-xray and other ones after links to the config pages
I thought those still worked, but with a different name?
oh you can't link to section headers? might want to have that too
what do you mean by this?
Basically, you should still be able to link to sections. Like the anti X-ray docs linked to the specific section in the world config, but you removed that
#204 doesn't format the code properly (the ifs), i'm also wondering if this could be the place where the tick temporal unit is shown?
Not really sure how big of a use that class is
I think that would make more sense if we added overload to the scheduler methods that took java's Duration
or date times maybe? idk if the Tick class can be easily added to those
Didn’t I fix that?
looks like only one, whoops
that's quite useful
Links to non-leaf nodes by clicking them
i think its pretty clear now whats a leaf node and what isnt @echo canyon
The whole doc uses 4 spaces except the custom inventory holder page.
wondering who would PR the only sane indentation character
(64c8c517cdf07f25f6ac9fd5) // @mortal yew (@me_adamrian / 1025568769092890625) has been banned by @neon atlas (202850073812402177)
Reason: invite to nsfw server spam
can someone with vercel access attempt that redeploy
or just, take a look at the dashboard to work out what is happening
They build just fine for me
What exactly @neon epoch
ollie breaking docs smh
Yeah the CI hasn't failed it didn't even start.
what the hell vercel, this is blatant favouritism
Ollie stop sending me emails smh
Idk how people can deal with GitHub sending email notifications
I would get way too many
The online notifications are enough
Because you're a maintainer lol
i only have a few
and you probably get info about everything
I only get stuff I interact with or got mentioned on
Have all GitHub notification mails deactivated. For private gitlab instance I need to deactivate them, was to lazy.
damm @brazen moon didn’t realise the docs were so out of date
I’ll get to all these updates tonight, thanks for going through them all
yes 😅, should be the last review
hopefully we can move to an external file definition on the site which can be mostly auto generated by running the server once
once we add the comments into the config files in Paper (when configurate merges that yaml comment pr)
then syncing it will just be run the server, copy the file over
Moving the data out is a good idea
it pretty much already is, just defined inline in a big block in that PR
super simple to have that be read from an external file
doesn't need to be now
Ah didn't actually look at the code earlier
Since this channel also has website in it's name: looking at that now, mostly to update the deps
depending on how automatic we want it to be... we could "host" this external file somewhere, and during a build process on teamcity, you start and stop the server somewhere to create that file which then gets updated where its hosted
then that file is fetched on the docs page, so literally never again would you even have to manually update the docs page
but again, not required for this reformat. just thinking about possibilities
lol I opened the old website project and was confused why there are so few deps
Yeah, depends on what we make the configurate comments look like as well because the docs site supports markdown but if the comments are saved in the yml file, that could get quite bulky / weird looking? I’m not sure about what the configurate comments will look like i haven’t been keeping up
I mean configurate comments are just comments we can set on each node essentially, but markdown obv doesn't render cause its just a plain text file
that’s what i mean, it could look a bit jank
How much is markdown used within the comments now? could just not do it
Few lists, few bolds, few hyperlinks. Some of the comments (especially the anti-xray ones) get really long though so could make editing the file really cumbersome
in all honestly, we could limit comments on some of the nodes when actually in the file, either way, defining the comments in the actual file for just the docs site is good
so like those comments might just not even need to be output to the yaml file during normal operations
What, so like a file comment and then a website comment which can be different but defaults to the file comment if web isn’t specified? Then, i guess these artifacts could be hosted on bibliothek or some other CI platform yall use
That could work, even if that wasn’t what you meant
some gradle task that generates the docs yaml
It would simplify the addition of config options though. as then you don’t have to make a seperate docs PR
it could just be a whole separate file in the paper repo, but then one of the test tasks make sure an entry for a website description exists for all docs
so the validation checks would fail if you didn't add an entry. I don't think I like the idea of defining both comments in the config java file, that would get too crowded
there are ways to do it that are nice and automatic
sure
Ok, think i got them all
@echo canyon I am just thinking about comments like this could get really annoying to put in Paper:
yeah, I think I like my idea of having an external file in Paper where to define long descriptions in markdown
then a validation check fails if there is a missing one
then in the paper source, you just put small one-line comments only for the yaml file
I like that idea too
<@&748618676189528155>
(64cc33e2cdf07f25f6ac9fdc) // @north ginkgo (@semen4091 / 582567774707253259) has been banned by @opal flare (105923848263753728)
Reason: porn ad
(64cc33e4cdf07f25f6ac9fdd) // @rugged prawn (@bretomart / 665981847490134018) has been banned by @neon atlas (202850073812402177)
lmao
how could you do this lynx
Note: Brandon died on the way back to his home planet. This is being resolved.
<@&748618676189528155> another one
@neon atlas now is your chance
There were stupid people, there are, and will be, not really Discord's fault this time
not very helpful nacio ?!?!?!
<@&748618676189528155>
attempt number 3
@neon atlas Ban brandon again ? xD
You're next.
But why mbax ?
@neon epoch anything missing from the docs PR btw ? like open reviews etc ?
given the PRathon coming up, I'd like to merge it
they all need reviewing and (some) merging, i was gonna ask for merge permissions on the docs repo but then i got sidetracked by valorant 

okay yea nice, I'll put that on my list for tomorrow
(merge permissions on the docs repo can easily be achieved by become triage
)
yeah idk how one does that lol 😄
#203 and #205 can probably be instamerged and then all the others probably need a review. LMK when you start looking at them, if its in the evening I may be around to help otherwise sunday ill be around all day most likely
Yea I'll give you a ping when we start 👍

https://github.com/PaperMC/docs/pull/209 Cursed github UI's be like
needs more commits
at least they are verified !
<@&748618676189528155>
(64cf92cccdf07f25f6ac9fe2) // @prime jolt (@bhargavshanti / 576457570483568642) has been banned by @neon atlas (202850073812402177)
Reason: Crypto scam
@neon atlas community manager when
ssssh don't say it out loud, mbax cannot know I am coming for his position

Hello
Hello
@neon epoch what happened with scheduler docs PR 🤣
hehehe
got a little bit trigger happy with the "delete branch" buttons
honestly thought it had been merged until i saw your ping lol
Thoughts on #paper-help message being changed over to a docs page? It can kinda explain Paper’s philosophy on “vanilla bugs” and what not, and mention those settings
Sounds sane
Ok, I have also finalised the versioning thing and have added a description here
This grabs the versions of projects from api.papermc.io. There is still some logic required todo with when you are opening an old doc like 1.19 or 1.18. This requests every 5 minutes to keep them u...
(64d7b40acdf07f25f6ac9ffb) // @wheat flame (@phantomonksad / 530448263103512576) has been banned by @neon atlas (202850073812402177)
Reason: Invite to nsfw discord server spam
why do they always do it here
it's always here and #folia-help
how do we capitalize and refer to paperweight userdev
Paperweight Userdev, Userdev, userdev, paperweight userdev, Paperweight userdev
i capitalise paper but then lower case the rest
well how is it on the actual github repo
it’s just paperweight
everything is lowercase, but idk if thats just cause that's the convention for artifact ids on maven repos
the artifact id is what
like you use it with https://plugins.gradle.org/plugin/io.papermc.paperweight.userdev
Paperweight UserDev
Paperweight Userdev
Paperweight userdev
PaperWeight UserDev
could be paperweight-userdev
I think if all lower, it should have a hyphen then right?
probably yeah
anyone else not know what a mermaid diagram was until ollie's pr?
like obv I'd see that kind of diagram, but the name
don’t worry machine, i still can’t remember if it’s amondition or admonition
we even googled the meaning
i’ll give your pr another look over later MM
Im really not vibing with the all lower case title @echo canyon. Other than that and the comment I just left, All looks good
can I change that title but have the header be that?
sidebar_label: Paperweight Userdev
Inside of the triple -'s at the top
Why not just "Userdev"?
RE: of vs from
Here is ChatGPT's opinion:
I understand your perspective. Both versions of the sentence are grammatically correct, and choosing between them can often come down to personal preference and style. The choice between using "from enabling these flags" and "of enabling these flags" is a nuanced one.
Using "from enabling these flags" suggests a direct cause-and-effect relationship between enabling the flags and the undesirable consequences. On the other hand, using "of enabling these flags" implies a more general association between the flags and the consequences, without explicitly stating a direct cause.
In formal writing, both versions are acceptable. If you prefer the version with "has" and "from," you can absolutely use it. The most important thing is that your sentence clearly conveys the intended meaning and is easily understandable to your audience.
@echo canyon
So it seems like an either or
genuine chatgpt usage moment
😄
This seems to suggest one is WAY more common. https://textranch.com/247218/consequence-of/or/consequence-from/amp/
consequence of or consequence from? - Which English form is more popular?
i think its cause the next word is a verb so it kinda fucks it a bit
i can rewrite it, thats just what aikar wrote
Consequences from seems redundant in some form.
Consequences implies “from” already I think.
“A consequence of laying in the sun is sunburn”
“A consequence from laying in the sun in sunburn”
“What are the consequences of your action? The consequences of my action…”
ChatGPT doesn't have opinions. It generates text based on text it has seen before. Sometimes it makes correct statements. Sometimes it makes incorrect statements. There is no way to know without doing the research you were trying to replace with ChatGPT.
Well a lot of what you said can be said about human opinions too…
Sometimes they are correct, sometimes they aren’t, gotta do more research to figure it out
ChatGPT doesn't do critical thinking, it's more like an advanced autocomplete. 🤷
of course
I was just saying that those reasons you gave also apply to human opinions.
paperDev userWeight
its not something thats specific to the pr, but i noticed it when looking at the version changer
@neon epoch
basically, when clicking on the version changer dropdown, it takes you to the root of the documentation, compared to changing documentation version, which keeps the current page
its not obviously versioned on mobile too, as the version selector is top level, despite the version being per-project
again not pr-specific
also clicking on the paper top level dropdown menu resets the selected verison back to latest
this is worse on mobile as to get the version selector, you need to go back to top level menu
🚰
something to do with DropdownNavbarItem.tsx - the props.to attribute
Well i cocked that up by not squashing, but nice
does it even matter really
kinda makes the commit log ugly cause there are random fix 2 things commits
Yeah i guess, if you can that would be great. it’s what i get for using mobile
generally I would name commits differnetly if they were in the full context of ALL commits, not just commits on a branch
yeah that makes sense 👍
it’s hard to follow because of how typescript works and it’s not like lovely java where all the imports actually link up so you can see usages
Eh it’s fine, whatever. Idk if it’s gonna break something, and I don’t wanna deal with it breaking if it does. Also don’t know if anyone is around who could cause I asked in another channel
looking here, https://docusaurus.io/docs it seems to occur on the docusaurus docs, so isnt a paper-docs issue?
That should just happen if the slug has changed i think
I was thinking for common "types" like duration, it might be good to have a legend/key with types to what they accept that could then be attached to each setting
to avoid duplicatating describing what a duration does
I did think that, but maybe something to come in the future. Also, on another note, i would be interested in maybe having the duration object support 2h30m
yeah, we could expand that for sure
seems useful and more readable than 150m
maybe in my configurate update I cna do that
Well, I was gonna say, we don't have a sane duration thing
though, do we even do duration stuff often? afaik there was only 1 thing for it
but, having a generally sane duration parsing thing is always nice
there are now 4 durations, all in the world config
3 for loot table stuff, one for chunk something
ah
chunk unload delays if my memory from 5 mins ago serves
would be nice if the restrict-player-reloot and restrict-player-reloot-time could be merged into one. Like almost a tristate of:
restrict-player-reloot: false | true | duration
there may be some edge case im not thinking of though
That would just create another issue with nether-damage-height a while ago if someone downgrade, it will break
But I guess we’ve already strongly adviced against downgrading anyway.
oh yeah the config migration system does not support downgrading
I mean I have zero issue with that. Just mentioning it.
If we never break anything, we will end up like spigot in 2069 and still support legacy
Can I get a review on 204 please anyone that reads this channel
thanks 
uuid or weakref
Good question, can do either
uuid is probably simpler though
Probably
public class MyConsumerTask implements Consumer<BukkitTask> {
private final UUID entityId;
public MyConsumerTask(UUID uuid) {
this.entityId = uuid;
}
@Override
public void accept(BukkitTask task) {
Entity entity = Bukkit.getServer().getEntity(this.entityId);
if (entity instanceof LivingEntity livingEntity) {
livingEntity.addPotionEffect(new PotionEffect(PotionEffectType.SPEED, 20, 1));
}
task.cancel(); // The entity is no longer valid, there's no point in continuing to run this task
}
}
Like this right @neon atlas
you are missing an else

thanks 🥰
This guide is designed for Non-Folia Bukkit Servers. If you are using Folia
the non-folia bukkit part sounds kinda weird to me
Sounds fine? I kinda get what you mean
just my opinion
you are wrong 😏
wow scheduler was actually merged 
yup, saw someone link the vercel in paper dev earlier so decided it was probably time 😄
ok, i have also "finished" the versioning PR -> see here for the comment that explains it
Hello, why when I import Paper on my Maven project does it not work?
sorry
ok thanks
I think we need to come up with standards for point of view for the docs
I don't really think there should be any (or very little) 1st or 2nd person point of view sentences
||I think you're a poo poo pee pee head||
but, yea, docs should generally be written from a "matter of fact" sense
So like no I, you, we, us, our, etc.
I'd kinda be inclined to agree, but, idk; stuff like that gets a bit easiest when you only have a bunch of people who are natives or going anal over enforcing it, idk
eh, when writing "your plugin.yml should look like this:" i dont really think there is a problem with that
maybe a github action checking for britishness might be worth it just cause me lol
We took over the world before, we can do it again
shhhh you arent supposed to reveal our plan
See, I get the disire to remove 1st/2nd person
1st person does start making stuff opinionative
if you can get a grammarly github action that would be cool
Now, some guides are just going to be that, but, if there is an opinion that wants to be passed, it should be passed in such a manner
yep, makes sense
your plugin.yml should look like this:
should be
the plugin.yml is formatted like:
urgh, idk the first feels more natural
first feels like a conversation
@neon epoch was gonna request changes rn
I'd stick with MiniMessage not minimessage
Ah Ok - i didn’t realise. Got used to all the other paper stuff being lowercase. Anyone lurking here can fix that if they want
my bad
Well, i can
I mean I can merge too so probably
i mean, i merged it 
i genuinely thought it was called minimessage tho lol
I'm starting to see the value in multiple people reviewing stuff
earlier it was just a thing slowing down a merge for me 
nah it is good, i don’t merge new pages until they have at least 2 reviews
I mean I did elsewhere
if it can be solved with like a simple check "is the inventory creatable by createInventory" I'd lean into just adding that in the examples and be done
also Malf couldn't reproduce so idk
i didn’t realise we had a docs page on that to be honest because i’ve always understood it was a bit of a grey area with regards to whether that was abuse of API
I mean I'm not closing that issue since I wrote the PR so kind of a conflict of interest
sure, i just don’t know enough about the inventory system to comment here hence i have left it open
and cat was speaking in tongues
I am Mini and I approve this message
I approve of your approval
I approve Nacio's approval of MiniDigger | Martin's approval
Found a bug
https://papermc.io/downloads/all
If you hover over Software the Changelog text will be still infront of the list
There is an open issue
If you can figure out how to fix it, be my guess
Z-index 9999 doesn't fix it so I am at the end of my skills 
- Delete:
.w-full .relativefromhtml.lpyjjhm.idc0_348 body div#__next main.flex-1 div.flex.flex-col.h-screen div.flex-1.flex.flex-row.min-h-0 div.flex-1.overflow-auto table.w-full.relativeso basically the table. - Delete
z-50fromhtml.lpyjjhm.idc0_348 body div#__next main.flex-1 div.flex.flex-col.h-screen div.flex-1.flex.flex-row.min-h-0 div.flex-1.overflow-auto table.w-full.relative thead.sticky.top-0.z-50.bg-background-light-10.dark:bg-background-dark-90.shadow-sm
Fixed it for me.
html body div#__next main.flex-1 div.flex.flex-col.h-screen div.flex-1.flex.flex-row.min-h-0 div.flex-1.overflow-auto table.w-full.relative
adding class z-1 to the table fixed it for me
Mind adding that here?
Ah I guess that creates a proper stacking context
anything z index < 50 works
That's the less intrusive solution.
Try isolation: isolate
just realised i might be doing work for no reason, would there be intrest in a pr that adds descriptions for spigot.yml / bukkit.yml etc to the modern docs format?
there is a level of "it's kinda wanted as we want to replace all documantation, but that stuff is eventually going to be merged into the paper configs"
but that stuff is eventually going to be merged into the paper configs
I mean you can cut and paste then
alr. i guess ill finish spigot.yml since ive started that, and we can go from their maybe
mostly im just mirroring the descriptions from spigots wiki, but i think some of it is outdated, so ill have to go into the source to see how it is actually used
you'd need to double check that their license lines up so you can do that
oh true
works here
I get 404
fine for me
Skill issue (loads)
I mean after you click Visit
someone turned on a feature requiring accounts
idk if something changed there
cat only you have an account
I mean not only you, but few people have accounts
Failed to authenticate. Please try again.
er
clicking logs or open group 404's for me, but otherwise is fine
and it says public
If I click through Assigning Domains it opens https://docs-git-fork-astei-main-1-papermc.vercel.app/, is this it or
yeah
fine then, ty
the "view deployment" button on github also seems to work fine
bleh @radiant summit i made the same mistake but make sure you squash and merge
someone is a fan of merge commits @opal flare
I am
it gives a more accurate history, and for something like docs it doesn't really matter
Paper repo is a different story
I was able to reproduce this on a test project, opened an issue with vercel. hopefully it'll be fixed soon, they're usually pretty responsive
ok i guess this means i cant do spigot.yml at all now, since im infected with the licensed stuff in my brain
Hi
The wiki content is not licenced for external use.
You can write stuff yourself even though you've read the wiki
The configs work the exact same whether the info about them is on Spigot's wiki or Paper's docs
so there will be some inevitable overlap
yeah true
i guess this also means i should do it with a copy of the paper source open in idea as well so it can be even better
And we're there to verify you're not breaking anything and Spigot can open an issue or message admins if they don't like something 
Looking for some feedback regarding this table comparing velocity with other proxies: https://github.com/PaperMC/docs/pull/222
Maybe someone can think of something else that Velocity does better, or has some ideas on how to word the entries in the table better
maybe add a note about the performance?
Didn't really want to put a blanket "Better Performance" claim there tbh
Maybe something about it being more efficient, e.g. less hardware required for more players. but idk how to say that 
i think there used to be a straight-up performance comparison on the website
but this was back in the velocity discord days
looking for it now
Hmmm, the spigot.yml has been copied into that PR - is that covered by GPL?
also, nice idea @serene vault but i think the content just needs tweaking a tad
Yeah I honestly wasn't too sure what to put there, just kinda went through the existing comparison pages
and the part about "modern API" is very vague but I also really don't want to list everything in detail there, then we have the same length as the text again 
now there isnt any header left, and once i finish all the options, the format would be differnt (due to replacing all the leaf options).
lmk if this is still going to be an issue
https://github.com/PaperMC/docs/pull/221/commits/b44267142a12334390e35420a9ee33b40f3f4fc9
i’m no lawyer so idk
next time i have a few thousand pounds lying around i’ll hire one to look into this
Maybe would be good idea run some benchmarking (for eg. using lots of bots) with memory/cpu usage? It wouldn't bee that reliabale, but it's still some point of reference
i tink it’s fine though as we are changing it a lot for documentation purposes
Bleh, i think it could be out of the scope for this.
problem is how to fit that into a small cell of a table
idea is to have like a simple overview
Like you can do 2 rows - Memory usage an CPU usage with * under table explaining how it was measured and on what device, java version etc.
i think that would be better as just a separate section all together
quick quick https://github.com/PaperMC/docs/pull/224
That was very quick
thank you nmf 
errm, you committed a localhost link in that docs pr, i would fix it but im only holiday. also, are we really recommending chatgpt?? Also, Todo Comments in teh code seems weird to me. just some things
It's genuinely amazing for these kind of language questions
TODO comments are intentional so there is absolutely no chance you will miss them when copy pasting everything
but yeah small things can be cleaned up still, just wanted it out in general for the announcement
Huh what's up with the paper-plugins page being under getting-started in the url, but not the other ones in that category
probably wrong as well, they all need cleanups cause they get moved around
np
<@&748618676189528155>
(64e50c59cdf07f25f6aca019) // @normal fjord (@yuliss03 / 869936375057752134) has been banned by @fair river (177150983258767360)
Reason: nsfw soam
(64e50c5ccdf07f25f6aca01a) // @normal fjord (@yuliss03 / 869936375057752134) has been banned by @ancient quiver (1098722699116810331)
Reason: nsfw spam
damnit
Rekt
Heh
soam
common API traps?
thinking stuff like the PIE event firing twice
and ik theres some other odd areas where "unexpected behavior is the default"
Also mentioning that path that the various config files are located at is probably useful. I can't see anywhere in the global config page that mentions config/paper-global.yml
Might be confusing, esp if you are expecting top level paper yml
interact event too!
@neon epoch thoughts on 225
it should be updated really, i have asked internally if we are recommending the snapshot as it is not marked as experimental
might just be a better idea to release 3.2.0
i have asked internally
same
what secret channel have you got 👀
The super secret contributors channel

@radiant summit needs bumping to snapshot
theres like a few things for velocity we wanted to deal with before we did a release
and then we can go break stuff
but, ngl, i have no idea what the current short/long term plan is regarding that
yeah then I'm just pushing
berate me later
no objections? great
dw, cursing you under my breath right now
No, just natural state
small change, fine as long as you ain’t breaking shit
ekhem
cough cough
is this empty line intentional? https://i.imgur.com/b5YpKN4.png
no, I do that alot apparently
i’m make concerned about the double newline above
Does anyone want to review https://github.com/PaperMC/docs/pull/221
Isn't it still a draft
like, open conversations with comments that start with "todo"
Yes, I can see that too
draft == WIP
fwiw we were supposed to check against Spigot since they didn't allow copy-pasting their wiki
i took a look, i need to cross reference this with spigots wiki and also there are a tonne of missing capital letters and whatnot
also just inconsistent format when compared to the yml in the other configs
spigots wiki is not licensed so we have to be careful
@jagged pecan ^ so that you are aware. I’ll take a proper look at the PR on sunday as i’m on holiday currently
yeah i re-wrote it from scratch. using only the paper repo with applyPatches
i was planning on going over the format later i think
request a review from me when you are done and i’ll take a look when i can
also @last bear regarding one of your comments
(64e8cf0bcdf07f25f6aca024) // @winter hedge (@oblachko_wc / 1059017518397800458) has been banned by @eager plover (126975485493248000)
Reason: Spam bot
is the example on http://docs.papermc.io/paper/reference/system-properties accturate?
is -Dpaper.log-level=FINE an actual property that can be set?
from the looks of it, no
YOONITS
Yea, I'll add it
beautiful 🤩🥰
Cool, still applies
i see more durations nowadays but to move over fully i would want it to support the multi unit format
maybe worth everything converting over at some point but it’s gonna be a PITA for converting configs over
I think for now we will keep this at ticks
Duration type needs some work to be more usable
exactly
it’s something i would like to see universally though. It’s just nicer :) - Sounds good
hopefully fixed, can you give it a review
ciao
Ollie, #227?
yeah i’ll look now
ive updated #221 to fix a lot of punctuation
cool beans, i’ll review in a bit
i dont think i can request an actual gh review unless: you have already reviewed it before OR i have push access
iirc thats just a quirk of github lol
Yeah np 👍
<@&748618676189528155>
(64ebb353cdf07f25f6aca027) // @vagrant path (@stevenissupacool / 954839129949106176) has been banned by @eager plover (126975485493248000)
Reason: Spam bot
your pr looks generally good @jagged pecan, im going through the file manually and changing the formatting to be inline with the previous pages. ill post it here when im done
👍
the quoting has removed some newlines
should i re-add the newlines with more enters?
is that for all affected newlines, or certain cases only? is there anywhere it makes sense to let it wrap, rather than me trying to manually keep it short enough
personally, i would just let it wrap there
looking at some other config files, the preference looks to be normal wrapping. ill only add newlines in places that really need it 👍
ive updated it on github, i dont think i can re-request since you didnt submit one on gh
<@&748618676189528155>
(64ebc36ecdf07f25f6aca029) // @wintry knot (@.kim._ / 527421404874801193) has been banned by @violet bough (172586171057176576)
Reason: spam
@neon epoch does https://github.com/PaperMC/docs/pull/212 support replacing inside config files?
This grabs the versions of projects from api.papermc.io. There is still some logic required todo with when you are opening an old doc like 1.19 or 1.18. This requests every 5 minutes to keep them u...
yes
thoughts on me doing the same thing for bukkit.yml (just looking at implementation, and not any online docs)?
Sure
Thank you. i’ll try and review later
hey @echo canyon could you give your opinions on https://github.com/PaperMC/docs/pull/212 - whether it be a code review or just general feedback please 🥰
i take it that it isn’t worth making command docs until owen does his shit
there are 2 missing properties from https://docs.papermc.io/paper/reference/system-properties i think
paper.uselegacypluginloading
bundlerreadonly
i think 2 is a mojang one? not sure what its useful for
I'll give it a read tomorrow 👍
I mean ll the non js stuff is fine. no clue if the react components are as they should be cause I barely know react. THey look fine tho
Is there a way to access old docs view? If not please add a toggle back... (if not possible please add a proper working search button) Ctrl+f is broken on the new docs.
Can you elebrate the issue around search? it works fine for me.
Or let me use another example
When I click "See all 11 results" it shows API stuff
Improving the search is a known thing




same
