development

[icy shadow]

whereas the 2nd will return 0

[atomic trail]

Gotcha, is any faster than the other for checking if the uuid is in the table?

[river solstice]

well the scenario with count would be something like:

try (ResultSet rs = st.executeQuery()) {
  if (rs.next()) {
    Long count = rs.getLong(1);
    // do whatever with the count of rows
  }
}

the select would be

try (ResultSet rs = st.executeQuery()) {
  while (rs.next()) {
    // do whatever with the rows, ex. get the column value for each row
    String foo = rs.getString("bar");
  }
}

[atomic trail]

That makes sense, ty

[river solstice]

if you intend to use the rows, go with SELECT *, if you only do the checking if it exists, go with SELECT COUNT(*)

since count technically should be faster

[spiral prairie]

That speed difference does not matter

[atomic trail]

Yeah I mean if I use count, it will create a default set of data for the player and then run SELECT * after anyway so doesnt make much sense to use it seems

Alright, also I assume it's fetch size that's 0 if the uuid isn't there? ```java
if(rs.getFetchSize() == 0) {
// Data doesn't exist
createDefaultData(uuid);
}

[river solstice]

fetchSize is not what you think

[atomic trail]

oh

[river solstice]

fetchSize is the amount of records will be fetched by jdbc

in (very) short, it's generally related to memory management

[atomic trail]

Oh okay, I won't touch that lol

Seems like I just need to check next()

[river solstice]

well, in his case, probably no, it won't matter

but I wouldn't use it in my work database with hundreds of millions of rows

[icy shadow]

if you want to do an upsert i think you can do that in 1 query

[river solstice]

well, yes, he can

@atomic trail if your UUID is marked as a unique key, then you can just use INSERT ... ON DUPLICATE KEY UPDATE ... (refer to docs for usage)

[atomic trail]

I got this to setup default data so not sure if that would work

                if(!rs.next()) {
                    // Data doesn't exist
                    createDefaultData(uuid);
                }

[atomic trail]

I'm getting the error java.sql.SQLException: ResultSet closed for this but not sure how?

It's from the bottom line

[dense drift]

you are still trying to access the balance even if the result set is empty

[atomic trail]

This should create the data though https://paste.helpch.at/egikafahil.java

If it's empty

Ah the resultset isnt updated

[dense drift]

bingo

[atomic trail]

Does this update the resultset correctly? Currently testing but it's not woorking correctly. It runs like it's in a while loop or something lol, but might be something different

                ResultSet rs = ps.executeQuery();
                if(!rs.next()) {
                    LOGGER.error("Data doesn't exist");
                    // Data doesn't exist
                    createDefaultData(uuid);
                    rs = ps.executeQuery();
                }

[icy shadow]

🤨

[atomic trail]

Sec lol

Nvm this runs every tick or so for some reason... lol

    private void autoSave() {
        new BukkitRunnable() {
            @Override
            public void run() {
                LOGGER.error("Data has been autosaved");
                save(false);
            }
        }.runTaskTimerAsynchronously(orbitalTesting, 1, ConfigManager.getDataSaveInterval());
    }

I can fix this tho ofc, but ty

For some reason I'm getting this error, not sure why

[19:13:57 ERROR]: 
org.sqlite.SQLiteException: [SQLITE_BUSY]  The database file is locked (database is locked)
    at org.sqlite.core.DB.newSQLException(DB.java:1030) ~[sqlite-jdbc-3.36.0.3.jar:?]
    at org.sqlite.core.DB.newSQLException(DB.java:1042) ~[sqlite-jdbc-3.36.0.3.jar:?]
    at org.sqlite.core.DB.execute(DB.java:881) ~[sqlite-jdbc-3.36.0.3.jar:?]
    at org.sqlite.core.DB.executeUpdate(DB.java:922) ~[sqlite-jdbc-3.36.0.3.jar:?]
    at org.sqlite.jdbc3.JDBC3PreparedStatement.executeUpdate(JDBC3PreparedStatement.java:98) ~[sqlite-jdbc-3.36.0.3.jar:?]
    at net.foster.testing.orbital.data.Database.lambda$serializePlayerData$5(Database.java:150) ~[orbital-1.0-SNAPSHOT.jar:?]
    at java.util.concurrent.CompletableFuture$UniAccept.tryFire(CompletableFuture.java:718) ~[?:?]
    at java.util.concurrent.CompletableFuture$Completion.run(CompletableFuture.java:482) ~[?:?]
    at java.util.concurrent.ThreadPoolExecutor.runWorker(ThreadPoolExecutor.java:1136) ~[?:?]
    at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:635) ~[?:?]
    at java.lang.Thread.run(Thread.java:833) ~[?:?]

Line 150 is: ps.executeUpdate();

[icy shadow]

i suspect you're not closing your connections in some way

[atomic trail]

Probably true lol

How should they be properly closed?

[icy shadow]

by calling .close() when you're done with them

or even better, use try-with-resources

you should really be doing this with all the AutoClosables, i.e. the connection, statement, and resultset

[dusky harness]

Idk about connection tho

[atomic trail]

This is the code, ig I should just make a local variable for it

    public CompletableFuture<PreparedStatement> prepareStatement(String query) {
        return openConnection().thenApplyAsync(connection -> {
            try {
                return connection.prepareStatement(query);
            } catch (Exception e) {
                LOGGER.error("Error when preparing statement for query: " + query);
                throw new RuntimeException(e);
            }
        }, EXECUTOR);
    }

[dusky harness]

Bc u need that the whole time

Unless it's like a pool thing

Then idk how it works

I should learn how to use sql

[atomic trail]

Not sure how AutoClosables work

But sounds ideal yeah lol

[dusky harness]

Google try with resourxws

It auto clpses

[icy shadow]

theyre things can be closed with try-with-resources

but also yeah, i think you can get away with keeping the connection open long-term

as long as the other two are closed

[atomic trail]

But I can't use try-with-resources for this can I?

            try {
                PreparedStatement statement = connection.prepareStatement(query);
                connection.close();
                return statement;
            } catch (Exception e) {
                LOGGER.error("Error when preparing statement for query: " + query);
                throw new RuntimeException(e);
            }

Or well I probably can somehow just not sure how it would be in this case

[icy shadow]

of course you can

try(PreparedStatement ps = connection.prepareStatement(query)) { ... }

oh

actually yeah it's a bit tricky in this case

you really have 2 choices

1. dont close it in that method and just remember to close it whenver you use it elsewhere

2. have this method take a lambda of what to do with the statement, then just return void

[atomic trail]

Either way I have to close it wherever I call the method?

A bit annoying to do all the time, but yeah I can do that

[icy shadow]

yes

[icy shadow]

i have a nice little example of this here: ```java
public <T> Future<T> query(String query, @NotNull CheckedConsumer<PreparedStatement> initializer, @NotNull CheckedFunction1<ResultSet, T> process) {
return Future.of(() ->
try(var con = dataSource.getConnection();
var statement = con.prepareStatement(query)) {
initializer.accept(statement);
var results = statement.executeQuery();
return process.apply(results);
});
}

(adjusted from using vavr so it might not be syntactically correct but you get the point)

 public Future<WarzonePlayer> load(@NotNull UUID id) {
        return database.query(
                "SELECT * FROM players WHERE uuid = ?",
                statement -> statement.setString(1, id.toString()),
                resultSet -> {
                    if (!resultSet.next()) {
                        return new WarzonePlayerBuilder().setPlayerId(id).createWarzonePlayer();
                    }
                    return loadPlayerFromResultSet(resultSet);
                }
        );
    }
``` then you'd use it like this

replace Future with CompletableFuture

[atomic trail]

I am not sure what you mean by CheckedConsumer

Just CompletableFuture as well?

Or CheckedFunction

[icy shadow]

theyre vavr types, just replacements of the standard ones except they can throw exceptions

just replace with the normal things & add exception handling urself

or actually it's probably better to define them urself

otherwise all the SQLExceptions will be really annoying to deal with

[atomic trail]

"Standard ones"? As in PreparedStatement and ResultSet?

[icy shadow]

what? no

Function and Consumer

[atomic trail]

Ohhhh lol ofc...

[icy shadow]

but you will get annoying exceptions if you do that

[atomic trail]

I'll create replacements now yeah

[icy shadow]

https://github.com/knightzmc/mittenlib/tree/master/core/src/main/java/me/bristermitten/mittenlib/util/lambda feel free to copy these verbatim as long as you credit me :))

although they come with a bit of baggage

[minor summit]

https://github.com/knightzmc/mittenlib/blob/master/core/src/main/java/me/bristermitten/mittenlib/util/lambda/SafeRunnable.java
psst, AutoCloseable

I am guilty of having used AutoCloseable as a throwing Runnable

https://github.com/knightzmc/mittenlib/blob/master/core/src/main/java/me/bristermitten/mittenlib/util/lambda/SafeSupplier.java
Callable

[atomic trail]

Can't really see why this is better, how does it handle exceptions?

This for example https://github.com/knightzmc/mittenlib/blob/master/core/src/main/java/me/bristermitten/mittenlib/util/lambda/SafeFunction.java

[icy shadow]

by throws Exception;

it just means your lambdas can throw exceptions

it doesnt exactly "handle" them

[icy shadow]

thats gross

[atomic trail]

Any way to use this without Errors class?

And Result

[icy shadow]

yeah it's called removing the method lol

and just copy the sneakyThrow impl

you kind of need it

or just wrap in RuntimeException

but not a fan of that

but really you only need the functional interface

[atomic trail]

But I can remove the applyCatching method?

[icy shadow]

yes

[atomic trail]

Gotcha, what is it actually used for?

[icy shadow]

well

a Result is just a more functional way of handling exceptions

which im personally fond of

so those methods are really just for convenience

im not sure im ever using them

but results in general are very handy

[atomic trail]

Very nice, is it possible to use statement inside resultSet? This explains it more ig

The exception handling is actually so nice to have

[icy shadow]

erm

Why do you need to do that lol

you have the results already

[atomic trail]

Ohhh yeah I'm dumb lol, mb

[atomic trail]

I can't figure out what this CompletableFuture.of() is?

Okay I really don't understand what you did for the query method lol

[icy shadow]

of is just supplyAsync

generally when you’re doing functional stuff like this you can just compare signatures

look for a static function taking a lambda and returning a future

There aren’t many options

[icy shadow]

which bit in particular?

[atomic trail]

Oh I might understand now, like this right?

    public CompletableFuture<PreparedStatement> query(String query, @NotNull SafeConsumer<PreparedStatement> initializer, @NotNull SafeFunction<ResultSet, PreparedStatement> process) {
        return CompletableFuture.supplyAsync(() -> {
            try(Connection connection = Database.openConnection().get();
            PreparedStatement statement = connection.prepareStatement(query)) {
                initializer.consume(statement);
                ResultSet resultSet = statement.executeQuery();
                return process.apply(resultSet);
            } catch(Exception e) {
                LOGGER.error("Error when preparing statement for query: " + query);
                throw new RuntimeException();
            }
        }, EXECUTOR);
    }

[icy shadow]

looks about right

Would definitely put e into the runtimeexception though

[atomic trail]

Oh yeah lol, forgot that

[icy shadow]

Oh also

Put the resultset in a try-with-resources, forgot to convert that sorry

[atomic trail]

Can I make sure that the resultSet is run first?

Or well I need to check this

        this.query("INSERT OR IGNORE INTO " + OrbitalTesting.TABLE_NAME + " (uuid, balance) VALUES (?, ?)",
                statement -> {
                    statement.setString(1, uuid.toString());
                    statement.setInt(2, ConfigManager.getInt("start-money"));},
                resultSet -> {
                    if(!resultSet.next()) {
                        // Data doesn't exist
                        createDefaultData(uuid);
                    }
                    return null; //TODO: FIX
                }
        );

[atomic trail]

Ah okay, ty

[icy shadow]

Wdym?

[atomic trail]

Well if the resultSet is empty I want to create the default data

Hence this

                    if(!resultSet.next()) {
                        // Data doesn't exist
                        createDefaultData(uuid);
                    }

[icy shadow]

what exactly is createDefaultData doing?

Because as far as I know resultsets are immutable

[atomic trail]

Oh wait this is wrong lol, but this is createDefaultData

It was used like this before

        CompletableFuture<PreparedStatement> cfPs = query("SELECT * FROM " + OrbitalTesting.TABLE_NAME + " WHERE uuid = '" + uuid + "';");
        cfPs.thenAcceptAsync(ps -> {
            try {
                ResultSet rs = ps.executeQuery();
                if(!rs.next()) {
                    // Data doesn't exist
                    createDefaultData(uuid);
                    rs = ps.executeQuery();
                }

                double balance = rs.getDouble("balance");
                econ.getBalances().put(player, balance);
            } catch (Exception e) {
                LOGGER.error("", e);
            }
        }, EXECUTOR);

[icy shadow]

okay… so what do you want that method to return, if anything

[atomic trail]

Before I changed query

I just want to set a "template" for the players data basically

[icy shadow]

what you’re doing here is very odd

and not very efficient at all

[atomic trail]

Probably yeah lol

[icy shadow]

You’re doing 3 queries when you could do this whole thing in 1

Look into “SQLite upsert”

Or possibly 2

This isn’t exactly an upsert

[atomic trail]

I'm doing it in 2 right now

[icy shadow]

well no

You’re doing 3, since you’re executing the lookup twice

[atomic trail]

This is what I currently use for createDefaultData and deserializePlayerData https://paste.helpch.at/exogesokid.java

I think I'm just doing it twice, but not too sure

Also a bit confused what to return in the resultSet

[icy shadow]

that does seem like ur doing the same thing twice yeah

[icy shadow]

this is a bit of a gotcha

you dont wanna use query for inserts because there is no resultset

what i did was had a separate method CompletableFuture<Void> execute(String query, @NotNull SafeConsumer<PreparedStatement> initializer)

and then you just do statement.execute(); and return null

no resultset

[atomic trail]

Ahhh yeah that makes more sense to do, thanks.

And yeah thanks for all the help, it's beginning to make more sense lol

[icy shadow]

np np

we'll have you monadding in no time

[atomic trail]

monadding? But yeah definitely lol

[atomic trail]

I'd recommend you to check #1147714001082523648 or #1147713999694209104

[lyric gyro]

Okay i just didnt see it sorry im dumb 🤧

[atomic trail]

No worries lol

[lyric gyro]

Thanks man

[atomic trail]

Uhhh how do I avoid the data being added to a new column every time? There should only be one column for every uuid ofc

[spiral prairie]

Unique attributes? A key (i forgor the name)

[worn jasper]

uh, help? Using morphia https://paste.helpch.at/goqehuxane.rb

[stuck canopy]

primary key?

[dusky harness]

Is that class in your jar?

[worn jasper]

yes

[worn jasper]

not even the forums are helping 💀

[dusky harness]

if you're still getting the error then feel free to share a github repo that has this issue 🥲

¯_(ツ)_/¯

and if the project isn't public then u can make a minimal reproducible example

[worn jasper]

Yeah its not in github, will try to recreate it and will send it here tmr

[royal hedge]

You probably need to specify the class loader as ur plugins one

[hoary scarab]

Apparently the previous hash on bitcoin block chain is little endian...
000000000000000000034934f36ef2518fbe17b53922cddd6dc849d789f06bf3 from block 807,746

But the pool I'm querying gave me
89f06bf36dc849d73922cddd8fbe17b5f36ef251000349340000000000000000

I tried simply reversing the string which didn't work. I also tried splitting the string every 2 chars and reversing which didn't work either. I know I am probs missing something but can't figure out what.

[torpid raft]

it's reversed but in blocks of 8 characters

00000000 00000000 00034934 f36ef251 8fbe17b5 3922cddd 6dc849d7 89f06bf3
89f06bf3 6dc849d7 3922cddd 8fbe17b5 f36ef251 00034934 00000000 00000000

since it's hex i guess it's split into 32 bit blocks

[hoary scarab]

Whats the easiest way to check the char per block? (Incase it ever changes)

[torpid raft]

wdym check char per block

like comparing two 32bit blocks to see if they are the same?

[hoary scarab]

I think I read your message wrong... It should always be 8 chars per block correct?

[torpid raft]

yeah probably

[hoary scarab]

👍 and it worked. Thank you.

[torpid raft]

nice 😌

it might change if they ever change the hash algorithm they use

[hoary scarab]

I doubt it ever will.

[torpid raft]

yeah realistically it's fine

[hoary scarab]

If I use something other then bitcoins block chain it might.

[atomic trail]

Can someone help me with ACF please, I'm trying to make an optional target for a /balance command, but the target is always just the sender, here's the code I have atm

    @CommandAlias("balance|bal")
    @Syntax("[player] &e - Check balance of player (default is sender)")
    public void onBal(CommandSender sender, @Optional Player target) {
        if(target == null) {
            if(sender instanceof Player) {
                String balance = numberFormatUS.format(econ.getBalance((Player) sender));

                sender.sendMessage(
                        ChatColor.translateAlternateColorCodes('&',
                                Objects.requireNonNull(commandsSection.getString("balance.no-target"))
                                        .replace("%balance%", balance)));
            }
            return;
        }

        String balance = numberFormatUS.format(econ.getBalance(target));
        sender.sendMessage(
                ChatColor.translateAlternateColorCodes('&',
                        Objects.requireNonNull(commandsSection.getString("balance.target"))
                                .replace("%target_player%", target.getName())
                                .replace("%balance%", balance)));
    }

[icy shadow]

use OnlinePlayer instead of Player

iirc Player always resolves to the sender

[icy shadow]

CF is an example of a "monad" which is an abstraction you see a lot in functional programming languages. im delusional in trying to functional-pill as many people as possible

[atomic trail]

Ah yeah didn't know that was a thing, but shouldn't Player in this case be private, because there's a getter and it's final? That's just what I've learned at least

It should be effectively the same, but still

[atomic trail]

Ahhh I see lmao, I'm monadding in no time then lol

[icy shadow]

it should be yeah but some of ACF is weird and old

[icy shadow]

we shall see, if you're using Streams, Optionals, or CF's in java you're partly there already

[atomic trail]

I'm not really using Streams or Optionals I think haha

[icy shadow]

☹️

[river solstice]

no streams no optionals? L

[icy shadow]

[worn jasper]

that.... is a good point, will check if it's that.

can't find anything about setting the class loader though

[sterile hinge]

set the context class loader

before initializing the Mapper

[icy shadow]

Classic

[worn jasper]

is that the one method present in the Thread class?

should I set it, initialize it and then go back to the old class?

[sterile hinge]

yes and yes I guess

[worn jasper]

Thread t = new Thread(() -> {
  datastore.getMapper().mapPackage("me.mrafonso.bdiscord.database.entity");
});
t.setContextClassLoader(plugin.getClass().getClassLoader());
t.start();

would this work?

[sterile hinge]

just use the current thread

but you probably need to do it before, because it needs to happen before the mapper is created

[worn jasper]

this worked lol, but now I am getting this: https://paste.helpch.at/aqahifosec.md

lol

it still can't find my class

well apparently it's relative path

so I made it just "database.entity"

but still receive this message: [dev.morphia.Datastore] No classes have been mapped.

[atomic trail]

Is it possible to add a command to CommandHelp for ACF? So this

    @HelpCommand
    public void onHelp(CommandSender sender, CommandHelp help) {
        help.showHelp();
    }

Will also show this command for example

    @CommandAlias("setbalance|setbal")
    @Subcommand("target amount")
    @Syntax("<target> <amount> &e - Give target player money")
    @CommandPermission("orbital.operator")
    public void onSetBal(CommandSender sender, String[] args) {
        // Some code

[hazy nimbus]

tbh I suggest you go ask in their discord

prolly gonna get a better answer there

[atomic trail]

They dont have one I believe

[hazy nimbus]

They do

Do you want to send a link in DMs?

[atomic trail]

Yes please

[stuck canopy]

what Entity interface have the getEquipement method?

[hazy nimbus]

https://jd.papermc.io/paper/1.20/org/bukkit/entity/Mob.html#getEquipment()

[stuck canopy]

ah LivingEntity

[hazy nimbus]

No, Mob

Apparently LivingEntity too

wtf is the difference betwen them

oh lol

mob inherits from livingentity

I wonder why the javadocs suggest that it's directly declared in mob

[stuck canopy]

I think Mob is implementing LivingEntity

[hazy nimbus]

it is

[stuck canopy]

yea

[hazy nimbus]

I would send a screenshot if I could 💀

[proud pebble]

you get image perms once you hit tier 2

[hazy nimbus]

ikr

But if you take a look at the mob javadocs

you can see that the method getEquipment is not listed under "Methods inherited from interface org.bukkit.entity.LivingEntity"

so I assummed it was declared there

[dense drift]

It is listed

[hazy nimbus]

I'm looking here: https://jd.papermc.io/paper/1.20/org/bukkit/entity/Mob.html

[proud pebble]

:/

[grim oasis]

bad paper docs 🤷‍♂️

[hazy nimbus]

it's kinda stoopid

[proud pebble]

what the hell did paper do to break their javadocs

[grim oasis]

https://i.imgur.com/BgqXWDl.jpeg
says the description is copy but doesn't list it

[hazy nimbus]

Yup

I didn't look there

cuz I was lazy

[grim oasis]

very weird

[hazy nimbus]

tbh I'm kinda interested too

tho why do they still use the jdk 8 style javadocs?

[proud pebble]

spigot does too

idk what the newer javadocs look like

[hazy nimbus]

hm

I swear I saw some new fancy javadocs site somewhere

can't find it now tho

apparently I was just hallucinating 💀

[grim oasis]

maybe you're just designing the new javadocs in your dreams

[hazy nimbus]

perhaps

[proud pebble]

i wonder if the reason im getting clientside lag is cus im sending lots of section packets instead of sending chunk packets

also im trying to figure out if i should have a map of chunk then section instead of one map that has all the block positions

    ChunkX      ChunkZ      SectionY    SectionRelativeBlockPosition
Map<Integer,Map<Integer,Map<Integer,Map<Short,Material>>>>

Map<BlockPositon,Material>

ofcourse for the massive map i have ill end up making objects for their specific storages

like it would be more simple with just 2 values but the massive map of values would make lookup times alot shorter (i think idk)

[sterile hinge]

why not Map<ChunkKey, Section[]>

[dense drift]

A map with 3 other maps looks like a bad design D:

[hazy nimbus]

Maybe not exactly helpful, but try using pakkit

but yeah

what's the XY here?

[river solstice]

triple map nesting

this would better not appear in my job during code review

you would have some shit to fix

[dense drift]

But they are HashMaps, the code has to be very fast!!

[hazy nimbus]

Just wondering, even though this seems like a horrible design and XYproblem, wouldn't it be better to just have Map<Short,Material>[][][]?
And then just access it using something[chunkx][chunkz][sectiony]

but that's cursed either

[pulsar ferry]

[atomic trail]

Currently having issues with TriumphGui if someone can help. When this runs it prints the player correctly and no errors or anything, but the gui isn't opened

    public void openSelectionGui() {
        LOGGER.error("Selection gui has been opened for " + player.getName());
        selectionGui.open(player);
    }

selectionGui = guiManager.setupSelectionGui(this);

setupSelection: https://paste.helpch.at/pehemeqixo.csharp

Does anyone have an idea why?

[pulsar ferry]

We probably need more context

Hard to tell anything is wrong from what you sent

[atomic trail]

Ah yeah I should probably have told you this. It works when I first join and use it (after server restart). But if I then leave the server and join back, the gui wont open. It works again after server has been restarted but yeah

Not sure what can cause this

It's run from this command

    @Default
    @CommandAlias("kits|kit")
    public void onKitsMenuOpen(Player player) {
        BePlayer bePlayer = playerManager.getBePlayer(player);
        bePlayer.openSelectionGui();
    }

But the BePlayer is not null either or it would give an NPE so not sure what's wrong

[minor summit]

nothing, the method is overridden in the mob interface to change the nullability annotation

[spark skiff]

Why not use Table

[river solstice]

maps are O(1), 1 * 3 is still O(1) !!! big speed, fast!

[warm steppe]

are you by any chance saving player objects in player manager?

if so, try saving uuids or names

[icy shadow]

It should do that automatically? That’s the whole point of CommandHelp

[proud pebble]

if your wondering what im doing, im making clientside private player mines

the whole point is so that i dont have to have hundreds of player mines stored on disk and all of them can be at 0,0 which means less chunks loaded aswell

[dusky harness]

anticheats might not like that tho

🥲

[proud pebble]

yeah ive thought about that

hears a fix 🌈 No Anticheat 🌈

[dusky harness]

💀

[proud pebble]

i might beable to hook into an anticheat and disable some of the checks when they are in the mine world since flight would be enabled anyways

probs the phase checks

[cerulean birch]

doesnt grim have simulated per-player worlds or something, sounds like it might sort of work with your system

[proud pebble]

im sure ill get to that as development continues

[cerulean birch]

probably

[proud pebble]

tbh thinking about it i might not even have to store the materials in memory at all, since i wont be saving them as whats inside the mine doesnt matter since its all cosmetic anyways

[atomic trail]

It only shows subcommands for a commandalias, so it just shows nothing with that

[atomic trail]

Not really, I'm just saving the BePlayer instance which contains the uuid and player

[warm steppe]

yeah but if you relog, your player object has different id and the manager class can't find it anymore (yes/no?)

[icy shadow]

The uuid will be the same

Assuming that’s being used as a key

[hoary scarab]

How would I compare hash and target in java?
Python I can just do hash < target

Example;```
Hash: cdd88b5d653fd2eb19e773cca822d080a118f1d543e4a03f4277f43f6c4bcfbf
Target: 0000000000000000000532ae0000000000000000000000000000000000000000

[icy shadow]

I mean it depends on the type

But that’s just compareTo

[hoary scarab]

Ah thought I would have to convert to integer or something. I'll try that thanks

I think I'm misunderstanding the description...
Would I use hash.compareTo(target) > 0?

[icy shadow]

other way around

0 = equal, negative = less than, positive = greater than

[atomic trail]

Yeah it is

[signal grove]

any way i can register a listener class, but ignore certain listeners inside it if they dont exist in the current version?

or will i have to split it up into separate listener classes

[icy shadow]

almost definitely the latter

[signal grove]

ok, thanks

[hoary scarab]

👍

[signal grove]

just curious, what are you doing with hashes that requires you to check anything except the compareTo being 0

unless it's some sort of locality sensitive hash, the < or > is probably useless info

[torpid raft]

bitcoin stuff

[icy shadow]

Decrypting the blockchain

[signal grove]

xD

[hoary scarab]

^^^

[worn jasper]

Still haven't solved this issue, if anyone has any ideas... 👀

[hoary scarab]

Been editing a mining script for a bit but figured I would port it to Java. (Maybe even c++ eventually) having issues getting the hexing right though.

[icy shadow]

lol

You’re gonna make negative money

[hoary scarab]

I don't care about making money with it. Just want to do it

[atomic trail]

Isn't this singleton and could be made static? https://paste.helpch.at/agifibemow.typescript

[torpid raft]

minecraft plugin bitcoin miner 🤯

people can finally make money from their vanilla survival servers averaging 0.03 active players

[royal hedge]

https://morphia.dev/morphia/2.3/javadoc/dev/morphia/mapping/MapperOptions.Builder.html#classLoader(java.lang.ClassLoader)

also plugin.getClassLoader()

[dusky harness]

wait a minute
isn't that creating a new database connection every request?

Note: I have barely used SQL, but I thought you are supposed to keep a database connection open and reuse it

[atomic trail]

Yeah it is, but apparently it's as it should be

Brister Mitten helped a lot yesterday and this is what I ended up with

So I assume it's right

[dusky harness]

oh ok

[atomic trail]

It's because the database is local and not remote I believe

[dusky harness]

but ig it could be static ye

[pulsar ferry]

Just because it could doesn't mean it should

[dusky harness]

tru

[atomic trail]

Wait yeah making it static wouldnt even fix my original issue

[dusky harness]

oh that makes sense

[atomic trail]

Waittt I should only have one connection for writing it seems
Sqlite takes care of the file level locking. Many threads can read, one can write. The locks prevent more than one writing.

Is this not true for plugin development? https://touchlabblog.tumblr.com/post/24474398246/android-sqlite-locking

[worn jasper]

but where do I set those MapperOptions, can't find anything....

or wait

[icy shadow]

I have kinda said this a few times lol

But also yeah don’t make it static

You’re losing a lot of refactoring potential if you do

most notably being able to move this to an interface so you can make a dummy implementation for testing

[worn jasper]

does #containsAtLeast() check offhand too? or only inventory?

[royal hedge]

It checks getContents so if that does, yes

[worn jasper]

hmm okay

also, random question, any way to fix the inventory click event being canceled not being properly canceled? like, if you click enough times with a good timing, the item stays in the slot it's not supposed to...

and it's not visual bug

[royal hedge]

U sure?

[worn jasper]

yes, 100%

even with priority in highest

[pulsar ferry]

Are you in creative?

[worn jasper]

nop survival

[pulsar ferry]

Should not happen in survival
Which version?

[worn jasper]

1.20.1

you have to click a bunch, sometimes it works, sometimes it doesn't

[royal hedge]

I still think it's a ghost item

[worn jasper]

but it stays there

[pulsar ferry]

Most likely a ghost item, but even then if you're in survival it should get removed

[worn jasper]

kinda doubt

let me try it out with a flint and steel

if it's a ghost item

it shouldn't light anything, right?

lol def. not a ghost item

even loses durability

[royal hedge]

isnt durability calculated client side as well

[worn jasper]

unsure, but either way an item shouldn't function if it's a ghost item, no?

[royal hedge]

ye it wouldnt

[worn jasper]

once you try to interact with it, it usually disappears

so it def. "bypasses" canceling the event

question is if this is an issue on paper/spigot, minecraft, or myself?

but I am just checking if the item I am putting in offhand is similar to X and if so, canceling it

nothing much

[royal hedge]

i think this is whats happening:

1. u put the item in the slot
2. u picked it out of the slot
3. the server receives the #1 and puts it back in your cursor
4. the server receives the #2 and puts it back into the inventory

[worn jasper]

that does make sense but can I do something about it? lol

since I assume that relies on timing of things

this looks like something out of my control?

could be wrong ofc

[leaden sinew]

Send your code

[worn jasper]

@EventHandler(priority = EventPriority.HIGHEST)
public void onClick(InventoryClickEvent e) {
    if (e.getSlot() != 40) return;

    assert normalItem != null;
    assert compactItem != null;

    if (e.getCursor() != null && (e.getCursor().isSimilar(normalItem) || e.getCursor().isSimilar(compactItem))) {
        e.setCancelled(true);
    }
}

normalItem and compactItem being 2 different ItemStacks that are the items I don't want to allow

[pulsar ferry]

Assert

[leaden sinew]

Why do you have assert?

[worn jasper]

purely for testing, those items are being fetched from a config file on start

lol

(but they are not null for context)

[leaden sinew]

Assert doesn’t do anything

[pulsar ferry]

Btw highest priority is the lowest lol

[worn jasper]

IDE likes those, I dislike yellow errors lol

[worn jasper]

what?

.-.

now I am confused

[pulsar ferry]

Bukkit's naming is dumb, highest importance runs last, it's not really priority (imo) so lowest runs first

[worn jasper]

bruh

but can that be the cause though?

[leaden sinew]

Probably not

[pulsar ferry]

Probably not, just letting you know

[leaden sinew]

What happens if you just cancel slot 40 and don’t check items?

[worn jasper]

but yeah, quite useful to know that lol

[worn jasper]

let me test that

[royal hedge]

yeah velocity gets this right with their PostOrder

except i think it should be an int and PostOrder a utility class

[worn jasper]

lol

@EventHandler(priority = EventPriority.LOWEST)
public void onClick(InventoryClickEvent e) {
    if (e.getSlot() != 40) return;
    e.setCancelled(true);
}

it still bypasses it

[royal hedge]

are you able to keep track of what is supposed to be in the offhand?

[worn jasper]

wdym?

[worn jasper]

at this point nothing is supposed to be in it, since I cancel all clicks in the slot

[pulsar ferry]

What are you doing btw? Just open a custom inventory and spam it?

[worn jasper]

not even custom inventory

player inventory

I am canceling the click in the offhand slot

of the player inventory

and if you spam it, it still somehow bypasses it

and it's def. not a ghost item

any chances the event is being called so many times, paper/spigot can't handle them all? lol

[worn jasper]

or like sparky said

[leaden sinew]

Is it a localhost?

Also maybe try https://hub.spigotmc.org/javadocs/spigot/org/bukkit/event/player/PlayerItemHeldEvent.html

[leaden sinew]

Are you cancelling the drag event too?

Sometimes when you click, you actually drag the item

[worn jasper]

no.

[worn jasper]

I am not, will check that.

[worn jasper]

not about held item

[blazing pewter]

anyone here know how to port rendering code from fabric 1.19.4 to 1.20.1

[atomic trail]

So I should preferably just have one connection for writing? And multiple for reading?

[icy shadow]

You can probably get away with just one connection if you’re not doing loads of concurrent reads

[atomic trail]

But it is possible to use more than one connection? How?

By checking if anything is already writing before doing anything?

[worn jasper]

seems like this did the trick

[icy shadow]

yeah you could do this with a lock on a connection, either just with a synchronized block or a proper lock object

I think

I am far from an expert on concurrency so this might be a terrible idea

[atomic trail]

Just creating one connection for writing I think lol

[icy shadow]

That won’t prevent concurrent writes

necessarily

You need a mechanism to prevent multiple threads from using the connection at once

Assuming you have multiple threads

[icy shadow]

And this is what locks are for

[worn jasper]

Any ideas how to disable this log using morphia in an mc plugin?

Logger mongoLogger = Logger.getLogger("org.mongodb.driver");
mongoLogger.setLevel(java.util.logging.Level.WARNING);

tried this but it didn't work

[atomic trail]

Oh, but I should still use a seperate connection for writing though right? To not check the lock when reading

I havn't used a lock before so not sure if this is how it should be implemented, is this correct?

    public CompletableFuture<Void> update(String query, @NotNull SafeConsumer<PreparedStatement> initializer) {
        ReentrantReadWriteLock lock = new ReentrantReadWriteLock();
        Lock writeLock = lock.writeLock();
        return writeConnection.thenAcceptAsync(connection -> {
            writeLock.lock();
            try(PreparedStatement statement = connection.prepareStatement(query)) {
                initializer.consume(statement);
                statement.executeUpdate();
            } catch(Exception e) {
                LOGGER.error("Error when preparing statement for query: " + query, e);
                throw new RuntimeException();
            } finally {
                writeLock.unlock();
            }
        }, EXECUTOR);
    }

[wheat carbon]

what are you trying to do?

make sure only one thing is done at a time on a connection?

i've never used locks, I'd just use a single thread executor

idk what's better per say

[atomic trail]

Trying to not write to database from multiple threads at the same time

But yeah a single thread executor would work as well I guess

[atomic trail]

It's to fix this

[wheat carbon]

yeah what I'd do is use a single thread executor and get rid of the lock stuff

and then use the sync version of this method writeConnection.thenAcceptAsync

in the task u schedule in the executor

[atomic trail]

I prefer having to just use one executor though. Because when saving I want to check if tasks are still running etc.

[icy shadow]

The lock needs to be global

if you’re doing that

and since you can use multiple threads for reading, yeah it makes sense to have a multi thread executor still

Tbh it might be easier to just do synchronized with some object as the lock

[atomic trail]

How would that work? Not sure what you mean

[atomic trail]

Yeah, perfect

[icy shadow]

private final Object writeLock = new Object();

private Connection writeConnection = ...

void writeSomething() {
  synchronized(writeLock) {
     // do shit
  }
}

something like this

[atomic trail]

Ehhh not really sure what that would do tbh, I think I'm sticking with the Java Lock system, seems simple enough lol

Would it work more efficiently or?

[icy shadow]

Synchronised is a lock

Yes

It’s the most primitive lock

[atomic trail]

Ohhh so actually just

    CompletableFuture<Connection> writeConnection = new CompletableFuture<>();
    private final ReentrantReadWriteLock lock = new ReentrantReadWriteLock();

    private CompletableFuture<Void> update(String query, @NotNull SafeConsumer<PreparedStatement> initializer) {
        synchronized(lock) {
            return writeConnection.thenAcceptAsync(connection -> {
                try(PreparedStatement statement = connection.prepareStatement(query)) {
                    initializer.consume(statement);
                    statement.executeUpdate();
                } catch(Exception e) {
                    LOGGER.error("Error when preparing statement for query: " + query, e);
                    throw new RuntimeException();
                }
            }, EXECUTOR);
        }

[icy shadow]

I mean

That’ll work but

When I wrote Object I literally meant that

You don’t need a lock object AND synchronized

[atomic trail]

How can it just use an empty object though? Does it just store the information in that object?

[icy shadow]

it’s just a sort of marker, it can be anything. Object is usually used though

as far as I know it’s based on reference equality

It creates a lock marked with a reference to the object, then that gets compared

[atomic trail]

Can't I just use this then? Or well I guess I can, but would that make it slower?

[icy shadow]

I suppose so

If you lock on "this" then any other code which knows about your object might choose to lock on it. While it's unlikely to happen, it certainly could do - and could cause deadlocks, or just excessive locking.

Ok yeah maybe don’t do that

https://stackoverflow.com/a/3369303

[atomic trail]

Ahhh yeah I will just do it on Object lol

[sterile hinge]

note that monitors currently don't work well with virtual threads, and it's uncertain how it will be in future

likely doesn't matter in your case I guess

but then, you can also just synchronize on the instance itself/make the method synchronized

[icy shadow]

That’s weird, you’d think it would be a pretty trivial change

[atomic trail]

Virtual threads? And I assume the monitor is what's checking if the lock is locked or unlocked in the object?

[icy shadow]

Making the method synchronized would need a bit of refactoring since part of the method is running supplyAsync, you wouldn’t actually be synchronising the right thing

[sterile hinge]

good point, but that's still an issue in the current code

[atomic trail]

Shouldn't synchronise be inside the cf then?

        synchronized(lock) {
            return writeConnection.thenAcceptAsync(connection -> {
                try(PreparedStatement statement = connection.prepareStatement(query)) {
                    initializer.consume(statement);
                    statement.executeUpdate();
                } catch(Exception e) {
                    LOGGER.error("Error when preparing statement for query: " + query, e);
                    throw new RuntimeException();
                }
            }, EXECUTOR);
        }

[icy shadow]

uh

yes with a big asterisk

maybe

[sterile hinge]

I think it comes down to monitors being managed by the runtime/native code while virtual threads are largely managed by java code

[icy shadow]

i dont really like having a global CF variable

[icy shadow]

ah yeah that makes sense. surely you could expose some API for managing native monitors though, right?

[atomic trail]

As in like this

            return writeConnection.thenAcceptAsync(connection -> {
                synchronized(lock) {
                    try (PreparedStatement statement = connection.prepareStatement(query)) {
                        initializer.consume(statement);
                        statement.executeUpdate();
                    } catch (Exception e) {
                        LOGGER.error("Error when preparing statement for query: " + query, e);
                        throw new RuntimeException();
                    }
                }
            }, EXECUTOR);

[atomic trail]

I need a global connection though? I guess I could make a new cf for the global connect every time though yeah

[icy shadow]

yeah uh

[sterile hinge]

I think they want to explore ways to basically get rid of monitors/their current implementation, but that will take some time I guess

[atomic trail]

Sooo essentially I shouldn't use the monitors then and just use a lock? Lol

[icy shadow]

the way i see it you have 2 options:

1. make it global and just .get() it once, as long as you know the delay will only be on startup
2. make a new CF every time and lazily initialize some global variable

i think something like this will work fine but it's a bit of a PITA to do everywhere #development message

what you might wanna do is refactor it into some method like T withLockedConnection(Function<Connection, CompletableFuture<T>>) where you handle all the locking in there

[atomic trail]

Only using the locking for the update method so I think it's alright without refactoring it

[icy shadow]

good point

[atomic trail]

lazily initialize some global variable? Not sure what this means

[icy shadow]

dont create the connection until you actually need to

but only create it once

[atomic trail]

What does the waits in get actually mean? Does it block the thread it's on until it's completed?

[atomic trail]

Ah I see, I think I'll use the first solution though, seems more simple

Modified the update method

    private CompletableFuture<Void> update(String query, @NotNull SafeConsumer<PreparedStatement> initializer) {
        Lock writeLock = lock.writeLock();
        return CompletableFuture.runAsync(() -> {
            writeLock.lock();
            try(PreparedStatement statement = writeConnection.prepareStatement(query)) {
                initializer.consume(statement);
                statement.executeUpdate();
            } catch(Exception e) {
                LOGGER.error("Error when preparing statement for query: " + query, e);
                throw new RuntimeException();
            } finally {
                writeLock.unlock();
            }
        }, EXECUTOR);
    }

Reverted back to the Java Lock class, just to be safe

[icy shadow]

yes

[icy shadow]

theres nothing unsafe about the synchronized locks lol

but suit yourself

also fix RuntimeException

[atomic trail]

Wdym? Exception is printed in logger, just prefer it that way, for some reason it didnt print when having it in the throwable

[atomic trail]

Wasn't there a potential issue with the monitor? For virtual threads

[atomic trail]

Ohhhhh, I see

[icy shadow]

virtual threads dont actually exist yet

well

unless you're on a preview version of java

[atomic trail]

But this?

[icy shadow]

yeah i mean, that's a very far-future thing

unless you plan on switching to java 21+ AND switching to virtual threads

it's probably not something you need to worry about

[royal hedge]

it is

you should plan to switch to java 21 and virtual threads

[icy shadow]

im not sure thats necessarily true

particularly not in this context where you're doing a lot of IO

with kotlin coroutines you're actually recommended not to use coroutines for IO

in general yes, though

[minor summit]

i mean virtual threads aren't exactly useful for file IO since you can't do them ops in a non blocking way like with, say, sockets

it'll just spawn a new carrier thread

by "useful" I mean fully take advantage of them

[icy shadow]

^^

[worn jasper]

any help here?

[pulsar ferry]

Why does it matter? It's just logging once that the client is created

[worn jasper]

it spams the console which isn't ideal

[pure crater]

I've been figuring out ways to change the fps based on the task. In short, here is a simple step by step process of what I'm doing:

1. Store a previous variable for the current milliseconds
2. Create a while loop that runs forever until completion
3. Within that runnable, check if the time passed by subtracting previous milliseconds from current is less than 42. If so, do continue and wait for the number of milliseconds to pass.
4. If 42 milliseconds passed, call the paint function, which will grab the frame and display as necessary.

As for a code example:

private void paintLoop() {
    final long previous = System.currentTimeMillis();
    while (this.running.get()) { // checks for completion
      try {
        final long current = System.currentTimeMillis();
        final long elapsed = current - previous;
        if (elapsed < 42) { // this is around 24 fps
          continue;
        }
        this.paint();
      } catch (final IOException e) {
        this.neon.logConsole(e.getMessage());
      }
    }
  }

The reason why I chose 42 milliseconds is because this approximantly 24 frames per second, since there are 1000 milliseconds and 1000/24 is 41.6666 or almost 42 milliseconds for each frame to show.

The issue with this is that it assumes that frames will always be 24 fps. In reality, many video games and video players have fluctuating frames per second to adjust based on how fast it takes for the frame to be processed/painted. I've explored solutions like buffering frames into a queue, but I couldn't find anything. does anyone know some possible algorithms?

[pulsar ferry]

Holovid uses thread.sleep to make sure it plays at the correct framerate, if you wanna take a look https://github.com/Holovid/Holovid/blob/master/src/main/java/me/mattstudios/holovid/display/DisplayTask.java#L73

[hazy nimbus]

Haha never figured that out

Or

You could deploy a log filter

and filter it there

but that's just a workaround that could be exploited

[stuck canopy]

pretty sure there is a #setDamage method for that

yep

@lyric gyro https://hub.spigotmc.org/javadocs/spigot/org/bukkit/inventory/meta/Damageable.html#setDamage(int)

Damageable#setDamage

[open pilot]

how do i make a item show as being "enchanted" whilst not actually having any enchants

[dense drift]

You can not, you can enchant it and then use the HIDE_ENCHANTS item flag

[craggy zealot]

"INFO java.io.FileNotFoundException: ./testWorld/data/raids.dat (No such file or directory)"

im creating a world that will be deleted but after i delete it it gives me this

I do set auto save off

https://paste.helpch.at/azicamatiq.rb

full error message

[snow dragon]

I need some help anyone here

[sterile hinge]

Just ask your question

[neat pier]

New Reminder Set!

In 4 hours, 59 minutes, 59 seconds:

continue working

Requested by: jerry6131#0

[minor summit]

what's heavier?

[inland walrus]

Can you help me?))

there is an error in my code, I can't figure out what it is. the console is showing a null error.

code:
https://paste.helpch.at/wuribigotu.java

I will send you the error now

[inland walrus]

error:
https://paste.helpch.at/iwepapalem.css

[gilded imp]

what took you one hour

[dusky harness]

Lol

[river solstice]

lmao

[river solstice]

well the error clearly tells you the lines where to look at

[mellow pond]

Also don't use spigot

[warm steppe]

why isnt spigot like deprecated or smth?

[dense drift]

Because it is not the case? Paper still depends on it and it cant just be shut down

[hoary scarab]

Same reason bukkit isn't. People are keeping it updated and just making forks of it.

[icy shadow]

lol not a single soul has cared about bukkit for years

As far as I know it’s effectively it’s renamed spigot now at this point too

[hoary scarab]

I still see questions about it from time to time

[sterile hinge]

because people don't understand that they use spigot

[strange remnant]

i use spigot?

[icy shadow]

maybe you do

[river solstice]

cool

[hoary scarab]

I still use spigot 😉

[icy shadow]

good for you

[worn jasper]

imagine

[worn jasper]

quick question, if I am using JDA in an mc plugin, would it be wise to execute jda code async? (using bukkit's scheduler)

[spiral prairie]

yes

[dusky harness]

no

well

if you need the scheduler then sure

but JDA already has its own async system that you can use

[spiral prairie]

isnt jda async anyway with the queue shit?

[dusky harness]

yea

[dusky harness]

why you put yes if you know about it then 😐

[spiral prairie]

because you can use jda async

[dusky harness]

they meant handling async themselves using bukkit scheduler

[spiral prairie]

eh not really

[dusky harness]

they said

using bukkit's scheduler

[spiral prairie]

just said if they can execute code async there

not to replace the queue with complete and be done with it

[dusky harness]

async using bukkit's scheduler, not in scheduler

but anyways, even for non-JDA stuff I'd recommend not using the scheduler unless you need to use the game ticks

[spiral prairie]

why

[dusky harness]

since you'd have to wait for the next tick

[spiral prairie]

the scheduler is a good way for the server software to manage all tasks without you using java executors

because afaik that can cause a few issues but not sure

[dusky harness]

manage all tasks
if you want repeating tasks every x ticks, then yes I recommend using it

but if you just want to run async code in general I wouldn't

[spiral prairie]

and also async tasks

you have to remember that youre not the host and that the server is very sensitive to async stuff

but i could be talking bullshit :=

[dusky harness]

why is it very sensitive
also note that the bukkit scheduler also uses java executor

[dusky harness]

it might be nice for bukkit to have a regular async thread pool for plugins to share

that doesn't run on ticks

[spiral prairie]

paper pr xD

[dusky harness]

lol
since if u have like 30 plugins that all manage their own async, that's a lot of threads

[spiral prairie]

exactly

[dusky harness]

but bukkit doesn't have that unfortunately 🥲

[spiral prairie]

if every plugin made an async executor

[orchid mural]

Async tasks already use a thread pool

[dusky harness]

ik I meant that it'd be nice if there was like runAsyncTaskImmediately

since currently, you can only run it on the tick loop

¯_(ツ)_/¯

[pulsar ferry]

Well minecraft is tick based so it makes sense that async tasks would also be tick based

[dusky harness]

true but with the amount of plugins that a server can have, I wonder how many threads a server has 🤔

[pulsar ferry]

That is not how it works, async tasks are run on a threadpool it'll only make a new thread if it needs one, there is normally only a few threads

[graceful hedge]

think it could be as simple as just exposing the cached thread pool through bukkit scheduler :>

like with getMainThreadExecutor

[dusky harness]

ye this is what i mean

bukkit currently doesn't have that, so I wonder how many threads a server has since a lot of the times, devs dont use the scheduler

¯_(ツ)_/¯

[graceful hedge]

myea

well cached thread pool doesnt cache threads for that long of a time tho

with the default factory method at least

[dusky harness]

oh, does it "return" threads back to the os (?)?

o

[neat pier]

📋 Your paste: tiltedoomer

https://paste.helpch.at/wevugoyake

A member of staff has requested I move your message to a paste,
Most likely because it contains a config/error/code snippet.

[pulsar ferry]

Suddenly a wall of text

[dusky harness]

can you show like a screenshot of what it currently does and explain what it should do?

alr

?imgur

[neat pier]

FAQ Answer:

You won't be able to upload images here directly to avoid spam, so please use https://imgur.com/upload to upload images/screenshots.
You can also use a screenshot service like gyazo or jinx and post those links here.

Requested by: dkim19375 • imgur

[dusky harness]

https://imgur.com/a/sLur9dc

[worn jasper]

yeah, but all the thing that don't have #queue(), are those unable to be executed async or don't have any impact on performance?

[dusky harness]

hm, well you have windowWidth and windowHeight, so maybe you could use those variables to calculate the offset of where all the stuff should be

oh it has to adjust size too

hmm

[dusky harness]

likely don't have any impact

you can always try to see if there's anything in spark when testing your plugin

[minor summit]

https://jd.papermc.io/paper/1.20/io/papermc/paper/threadedregions/scheduler/AsyncScheduler.html :p

[dusky harness]

WOA

wow

[dusky harness]

nevermind no need 🙃

[graceful hedge]

ooo poggers

[wise nest]

part of my plugin involves forcing the player to use an end crystal item
how can i check that the player is able to place the end crystal item (like making sure that another plugin or spawn protection wont block it) before spawing the entity (i assume i have to just spawn in the entity manually)
i want it to be as if the player had just right clicked with the item

[cerulean birch]

is it at all possible to use hex colors in spigot's api without adding paper/minimessage, make with ChatColor or smthing

[dusky harness]

yes
I think it's like §x§a§b§c§d§e§f for #abcdef

[cerulean birch]

works, ty

[spiral mural]

ey, any github wizzards here? can't for some reason with my private access token generate a valid url that works for external access?

at this point I've enabled everything for the url to be accessable using the token :p

[dusky harness]

how are you testing the token?

[spiral mural]

I just want to be able to use a resource pack in a private repo

so by url

/x/x/raw/main/pack.zip?token=x so like

[dusky harness]

does github support ?token=x?

https://developer.github.com/changes/2019-11-05-deprecated-passwords-and-authorizations-api/#authenticating-using-query-parameters
Seems like it was deprecated in 2019... it might've been removed by now

https://docs.github.com/en/rest/overview/authenticating-to-the-rest-api?apiVersion=2022-11-28#about-authentication since in here they use headers

¯_(ツ)_/¯

I've never actually used github api before, this is all from a bunch of googling, so I may be incorrect on some parts

[spiral mural]

aww guess am using dropbox. appreciate the lookup!

[somber dew]

I am trying to change the text of a TextDisplay using the .setText() method. However, anything I put into that method becomes this:

text:'{"extra":[{"text":"This is a test"}],"text":""}'

I can not figure out why it does this, and the documentation says nothing of this.

Documentation:https://hub.spigotmc.org/javadocs/bukkit/org/bukkit/entity/TextDisplay.html#setText(java.lang.String)

[leaden sinew]

Can you send your code?

[somber dew]

if (command.getName().equalsIgnoreCase("createEntity")) {
            if (sender instanceof Player) {
                final Player player = (Player) sender;

                final TextDisplay textDisplay = player.getWorld().spawn(player.getLocation(), TextDisplay.class);
                textDisplay.setText("This is a test.");

                // Show a particular entity
                if (!entityHider.getMembership(player, textDisplay.getEntityId())) {
                    entityHider.setMembership(player, textDisplay.getEntityId(), true);
                }

            }
        }

[leaden sinew]

That’s weird, maybe it’s a spigot bug

[somber dew]

Great XD

Just made a ticket: https://hub.spigotmc.org/jira/browse/SPIGOT-7483

[dusty frost]

I mean that just looks like Components or similar

guess .setText() isn't the method you use to actually display stuff, idk what TextDisplay is ngl but something like that

[sterile hinge]

pretty sure that's just how spigot's text -> component parsing works for years

I'm almost certain this is also the case when sending messages to players etc

[somber dew]

If I wanted a player to be able to run my command, and pass in an entity, such as a TextDisplay, how would I parse a selector (@e[type=text_display])?

[sterile hinge]

Use a command library that provides functionality for that

[somber dew]

Do you know of one? I'm pretty unfamiliar with this stuff.

[sterile hinge]

brigadier, cloud, commandapi

[wooden yoke]

Hi, how can I disable expansion tasks when the expansion is disabled or reloaded?

[atomic trail]

The parameter for an enum is only called once right? And when is that?

    public enum MiscSlot {
        FILLER(ConfigManager.getList("guis.upgrade.misc.filler.slots"),
                Utils.deserialize(ConfigManager.getConfigurationSection("guis.upgrade.misc.filler.unlocked-item.item")),
                Utils.deserialize(ConfigManager.getConfigurationSection("guis.upgrade.misc.filler.locked-item.item"))),

[sterile hinge]

yes, you pass those values to the enum constructor when the class is initialized

[atomic trail]

Ummm when would an enum class be initialized though? Lol this is probably very obvious but havnt used enums much

[sterile hinge]

that's the fun part, basically on first access to anything of it, but that's something that shouldn't matter in the best case

[leaden sinew]

If you want to support editing config values without restarting the server I would not do that

[atomic trail]

Ohhh yeah that's true, but wouldn't it be kinda inefficient to do it another way?

[atomic trail]

Perfect! Thank you

[leaden sinew]

No, you should just store the path of the item in the enum, not the actual item

[atomic trail]

But then it would be less efficient every time I call it? I suppose it just depends on if I want it to be reloadable

[leaden sinew]

You can cache the values

[atomic trail]

How do you mean?

[leaden sinew]

class SomeCachedValues {
    private final Map<String, ItemStack> cached;

    public void load(FileConfiguration config) {
        this.cached.put(MiscSlot.FILLER.key(), config.get(MiscSclot.Filler.key());
    }
}

[sterile hinge]

you can use an EnumMap in such case

[leaden sinew]

It looks like they have mutliple keys per enum though

[sterile hinge]

you can use multiple EnumMaps in such case

[leaden sinew]

Good point

[atomic trail]

I think I'm fine with just not caching it and requiring a server restart lol

It's just for one server anyway

And performance is prioritised

[leaden sinew]

Enums really shouldn't be dynamic like that though

[atomic trail]

Why not?

[leaden sinew]

And if that's where your performance issues are, you have bigger issues

They're meant to be constant

[sterile hinge]

EnumMaps perform very well

[atomic trail]

They are constant in my case though no?

[leaden sinew]

They aren't constant at compile time though

Plus there's a lot of opportunities for it to throw errors, what if the configuration section doesn't exist, or what if the item it returns is null?

[proud pebble]

looks like a json representation of the text

weird

[sterile hinge]

lmao as I said

[proud pebble]

id assume theres a reason why its

extra {
  text:"This is a test"
}

instead of just

text:"This is a test"

probs something to do with how mojang made it

[wise nest]

in a plugin, if you perform an action which would call a listener (eg: player.attack could call entitydmagebyentity), will all the listeners be ran immediately, or will they be ran after the block of code (method) from the first plugin

i know that with callevent() it runs immediately but i dont know about this

[sterile hinge]

that depends on how it is implemented

[wise nest]

what do you mean by this?

it is just a normal bukkit api method

[hoary scarab]

They would run after your call, unless you ran it in a scheduled runnable

[sterile hinge]

some events are triggered on the next tick if some value was changed, some events are triggered directly I guess

[wooden yoke]

?

[hoary scarab]

Ah yeap. Read his message wrong... sirywell is right.

[atomic trail]

Will this loop through the tiers in the order 1, 2, 3 etc. all the time? for (String strTier : section.getConfigurationSection("tiers").getKeys(false)

[hoary scarab]

I think getKeys() is List so it should keep order

[sterile hinge]

it returns a Set, so no guaranteed order

[atomic trail]

Ah I see, alright

[hoary scarab]

Nvm it's a set.

Yeah what sirywell said.

[wise nest]

why does Minecraft sometimes just decide not to send a cancel digging packet

[atomic trail]

Yeah, which does make the most sense performance-wise ig

[sterile hinge]

I don't think it's a performance consideration, but more of the semantics of the config basically being nested maps

(and keys are a set obviously)

[atomic trail]

Not too familiar with CompletableFutures yet, what could be the reason that thenAcceptAsync doesnt run?

        CompletableFuture<BePlayer> completableFuture = database.deserializePlayerData(uuid);
        LOGGER.error("CompletableFuture created");
        completableFuture.thenAcceptAsync((BePlayer data) -> {
            LOGGER.error("Inside CF");
            Bukkit.getScheduler().runTask(bossEvents, () -> {
                LOGGER.error("Inside bukkit task");
                bePlayers.put(data.getUniqueId(), data);
                LOGGER.error("Player added to bePlayers: " + data);
            });
        }, Database.EXECUTOR);

Only the first debug is printed

Is it because the data is null / void?

[dense drift]

It should run after the deserializePlayerData returns a value. There might be an error thrown, try to use the method that gives you the error too (Throwable)

[atomic trail]

I use BM's SafeFunction, not sure if it just removes the exception?
https://paste.helpch.at/kuvegoroya.typescript

sneakyThrow is just this

    /**
     * Sneakily throw a checked exception as an unchecked exception
     *
     * @param e   the exception to throw
     * @param <E> the type of the exception
     * @throws E the exception
     */
    public static <E extends Throwable> void sneakyThrow(Throwable e) throws E {
        //noinspection unchecked
        throw (E) e;
    }

@icy shadow Sorry for the ping if it's annoying, but since it's your code you probably know why?

Or well the function system is at least

[icy shadow]

CF's swallow exceptions

if one gets thrown then thenAcceptAsync wont run

[atomic trail]

Are they stored anywhere so I can print them to debug?

[icy shadow]

thats what .exceptionally or .whenComplete is for

[atomic trail]

Ohhh I see, so this would be correct usage it seems

        CompletableFuture<BePlayer> completableFuture = database.deserializePlayerData(uuid);
        completableFuture.thenAcceptAsync((BePlayer data) -> {
            Bukkit.getScheduler().runTask(bossEvents, () -> {
                bePlayers.put(data.getUniqueId(), data);
            });
        }, Database.EXECUTOR).whenCompleteAsync((input, exception) -> {
            if(exception != null) {
                LOGGER.error("Error when loading player data", exception);
            }
        }, Database.EXECUTOR);

[dense drift]

ig you can simply schedule the bukkit task on whenComplete ?

[atomic trail]

What's the difference between thenAccept and whenComplete then? Other than the exception ofc

[minor summit]

thenAccept(Async) only runs when the future completed successfully, if it completed abnormally (threw an exception) the ex is passed down the chain, so you can have a whenComplete/exceptionally further down that handles it
whenComplete(Async) runs when the future completes either successfully or abnormally, and it gives you the two potential results (value or error)

[icy shadow]

thenAccept is analogous to Optional#map, whereas whenComplete is more like Optional#orElse

[minor summit]

ifPresentOrElse

[icy shadow]

oh yeah i forgot thats a thing

yeah

also i dont think you really need to be doing .xAsync over and over again

[atomic trail]

Ahhhh gotcha, thank you for explaining! :))

[atomic trail]

Oh yeah that would make sense, I guess only the first time, and then it runs on that thread?

Any way I can test which thread it's currently running on?

[icy shadow]

Thread.currentThread()

but i gave an example of this a while ago

unless you explicitly override it, all the tasks will run on the thread that calls .complete() or equivalent

[atomic trail]

Oh yeah mb

Ty

(Ignore the async) Do I call the executor twice in this?

[icy shadow]

maybe

i mean

it certainly sends 2 tasks, yeah

it's not like it can magically compress them into 1 task

[atomic trail]

Wait if I call whenCompleteAsync I cannot use the executor, so it either doesn't register the task or it's just one long task? Or I'm just completely wrong lol

[icy shadow]

you dont need the second parameter if you're not using *Async

[icy shadow]

because it's defined as this

so yeah effectively it is 1 long task

[atomic trail]

I thought I should call async on the first "task", or is it on the last, so in this example whenCompleteAsync?

Not sure what .complete would be

[icy shadow]

neither is necessary

[icy shadow]

supplyAsync i guess

just use supplyAsync with the param and u dont need any extra asyncs

unless you want to switch threads

which is unlikely

[wise nest]

why does minecraft sometimes decide not to send a cancel digging packet?

[atomic trail]

Why that and not thenApply when making the query? Sorry for all the questions just trying to learn

[atomic trail]

First screenshot is where the cf for this is created

[icy shadow]

what is openConnection doing

i mean really it doesnt matter

but earlier is probably better

and you only really need to do it once

[atomic trail]

Creating a new connection

[icy shadow]

well obviously lol

um

[tardy cosmos]

I sure hope it is

[icy shadow]

why not just use supplyAsync here

[tardy cosmos]

this looks like a perfect candidate for supplyAsync

[icy shadow]

also i think that approach to handling exceptions is a bad idea

all that's gonna end up happening is your futures will randomly never complete, and all you get is a single logger message to diagnose

ALSO you are leaking connection objects by never closing them

[tardy cosmos]

Any recommendations for a small dependency injection framework for plugins (Java)? Looking at Feather right now basically cuz it's small. I haven't used DI frameworks in plugins before

[dusty frost]

Idk how small Dagger is but I quite enjoyed it when I used it

[icy shadow]

having to explicitly bind everything with dagger got kinda tedious in my experience

guice is very nice

[tardy cosmos]

I've used Dagger, I'm not a fan of how the project has to be compiled once for it to resolve though

[atomic trail]

So this would be an improvement I guess? And then I close the connection when I'm done with it manually?

Oh wait I dont need complete

[icy shadow]

yeah uh

[atomic trail]

Just this return DriverManager.getConnection("jdbc:sqlite:" + BossEvents.DATABASE_PATH);

[icy shadow]

stop doing the RuntimeException without params 😠 that gives ZERO useful information

[leaden sinew]

It’s a client bug, I forget the specific circumstance it happens, it may be after breaking a block

[atomic trail]

But it's better than returning null and I print it in the logger anyway? Lol

I've just had issues with printing it in the exception itself

[icy shadow]

honestly im not a fan of all the manual logging

[tardy cosmos]

that runtime exception black hole scares me

[atomic trail]

Why not?

[tardy cosmos]

Rethrowing it like that is fine just make sure the original exception can still bubble up or handle the future exceptionally

[icy shadow]

ideally you would not do any explicit try-catches, and then just log/handle any exceptions at the end in whenComplete

[tardy cosmos]

exceptions being eaten by the void is scary

[atomic trail]

Ohhh yeah forgot that the cf "stores" the exception or whatever

Or cause I suppose it would be

[leaden sinew]

Yeah but if there’s no exception printed, than it gets ride of the error

[tardy cosmos]

hiding errors is not getting rid of the errors

[icy shadow]

idk really this might just be me being opinionated. it's just far nicer if you can skip all the imperative jargon and then only do 1 error handling thing at the end

maybe thats too functional-pilled though

[leaden sinew]

If you can’t see it, it isn’t there

[tardy cosmos]

everyone should take the functional pill

[icy shadow]

[atomic trail]

Idk I feel like I save a few seconds or minutes when I debug it like that, but the exception probably explains it better yeah

[icy shadow]

i think the thing to consider

is that messages in exceptions are usually pretty clear already

[atomic trail]

Yeah true

[icy shadow]

you can do debug logging but is "Connection to database could not open" really adding much when the error will already be like SQLException: Connection Failed to <address>, timed out or something, plus a stacktrace?

[atomic trail]

But in this example, I wouldn't think right off that it would be because the connection couldn't open I think

[icy shadow]

what example?

[atomic trail]

Oh would that be the exception? Yeah.... that explains it pretty well ll

[icy shadow]

i mean probably

[tardy cosmos]

why is this creating a new future twice

[icy shadow]

it's not like it'll just be a random exception

theyre usually somewhat informative

[atomic trail]

Yeah, I guess I can add my own if I at some point is confused, but sql explains it well I'd say

[atomic trail]

Thank you...

[tardy cosmos]

that's better

[icy shadow]

mostly

[atomic trail]

Debug is removed

[tardy cosmos]

it should at least work now

[icy shadow]

the catch-all Exception is possibly a bit of a smell but

yeah

[tardy cosmos]

can I have you change that to only catching an SQL exception for my sanity

[icy shadow]

they're almost always informative enough. a few vague messages arent worth all the extra try-catch maintenance imo

[atomic trail]

But doesn't it do the exact same though?

Effectively

[tardy cosmos]

well yes but actually no

that could catch a divide by 0

[atomic trail]

Yes, but there is no divide by 0 in there, so it would just print the same exception anyway

[icy shadow]

how do you know

[tardy cosmos]

there still could be in getConnection and you'd never know

[icy shadow]

getConnection could throw anything

but

[atomic trail]

Because of this, so I'd also have to add the second one

[atomic trail]

But then I'd prefer it to also throw the divide by 0 exception imo

[icy shadow]

it will?

it'll still get thrown even if you dont catch it

the CF mechanism will handle it, and you'll get to gracefully handle it at the end of the pipeline

[tardy cosmos]

it would throw it and not get caught cause you don't generally want to catch a divide by 0

[atomic trail]

But wouldn't it also handle it if I catch it with just Exception e? Sorry if I'm sounding like I'm trying to discuss, just trying to understand

[icy shadow]

i mean yes and no

[tardy cosmos]

It would also handle it with Exception e yes, it would just also handle cases other than SQL exceptions

[icy shadow]

sure, you'll rethrow it as RuntimeException yeah

[tardy cosmos]

You generally only want to catch exceptions you're expecting

[icy shadow]

^^ that's the real point

[tardy cosmos]

Putting a catch all for Exception is a bit dangerous in that sense

Catching SQLException would handle all exceptions the getConnection method is expected to throw

[atomic trail]

So the errors in the end should really be unexpected exceptions kind of?

Idk if that makes sense

[tardy cosmos]

you can think of it like that I guess

[icy shadow]

i mean arent all exceptions unexpected?

but yeah

[tardy cosmos]

some are slightly more expected than others lol

[atomic trail]

Yeah not sure how else to explain it hahah

[tardy cosmos]

do whatever your heart desires

[atomic trail]

Nonono, that's horrible coding advice for me lol

Okay one last thing, then I'll leave you guys alone

How should the connection be closed properly in the end?

[icy shadow]

i would do that in query after everything else

in fact i vaguely remember the code i sent for query including the code to close the connection lol

but maybe im tripping

you can just put it in the twr and all should work

[atomic trail]

twr?

[atomic trail]

I don't think it did but probably lol

[tardy cosmos]

I personally have my connections auto close by requiring them to be used in a lambda
https://github.com/Rosewood-Development/RoseGarden/blob/master/src/main/java/dev/rosewood/rosegarden/database/MySQLConnector.java

[atomic trail]

But wouldn't it close before I can use it then? If it's before the return in openConnection method?

[icy shadow]

far from it

[icy shadow]

thats kinda like we're going for too

[icy shadow]

try with resources

which runs after the return

[atomic trail]

Not functional coding!

Lol sorry

[tardy cosmos]

at least it's only catching the SQL exception lel

[atomic trail]

True true

[atomic trail]

Ohhh yeah smart

[icy shadow]

and theyre not using CF so you dont have many other options 🤓

[tardy cosmos]

yeah my lack of CF here doesn't help you at all

you could add another stage at the end that closes the connection I guess, I don't know if there's a way to do that from inside the getConnection call though

[atomic trail]

twr like this I assume?

            try(Connection connection = DriverManager.getConnection("jdbc:sqlite:" + BossEvents.DATABASE_PATH)) {
                return connection;
            } catch(SQLException e) {
                throw new RuntimeException(e);
            }

[icy shadow]

um no

[tardy cosmos]

now that one would close the connection immediately

[icy shadow]

.

[atomic trail]

I thought that was what the twr would do

[icy shadow]

there is no way to do it inside getConnection without heavy refactors, yeah

[atomic trail]

Ohhh in query yeah... Sorry missread

The connection is in the cf though so not sure how twr would be used?

Nooo

[icy shadow]

same way you're already using it

[atomic trail]

ignore the exception

You didnt see anything

It isn't this is it? lol

[tardy cosmos]

openConnection thenApply thenRun and close it in there

Err maybe not

the thenApplyAsync is gonna eat the connection

add a finally block to the try/catch there and do connection.close()

[atomic trail]

It warned for throwing exception inside finally so I just did this, correct?

[icy shadow]

definitely at least log something on the error

closing might fail

[atomic trail]

Why not throw an exception though?

[icy shadow]

good question

that works too ig

[atomic trail]

IntelliJ just warned me about it for some reason

[tardy cosmos]

The newest version of intellij warns you if you simply do e.printStackTrace() at all lol

which I guess yeah they're right it should be logged elsewhere

[icy shadow]

oh wait inside finally

[minor summit]

suppressed exceptions moment

[icy shadow]

i would just

[tardy cosmos]

the one inside finally can still fail but just throw a wrapped runtime exception there I think

[icy shadow]

try( blah; Connection newConnection = connection) {
  do all
}```

no need for finally at all

[tardy cosmos]

oh does that just work

[icy shadow]

think so

[tardy cosmos]

I wasn't aware you didn't have to do the actual assignment in a twr

[icy shadow]

that is assignment

[tardy cosmos]

I guess,,

[atomic trail]

I mean no errors when compiling at least

[minor summit]

you don't even need assignment, you can just do

try (connection) {
  // ...
}

[atomic trail]

Not in java 8...

[minor summit]

sucks

[tardy cosmos]

java 8

java 21 gang

[minor summit]

t r u e

i need to update

[tardy cosmos]

I don't think it's out yet

a couple days I think?

[minor summit]

it released today

[tardy cosmos]

oh shit

[atomic trail]

Yes, the server runs java 8 so

[minor summit]

temurin isn't built tho

[tardy cosmos]

heck

do we think 1.21 will use Java 21

[icy shadow]

unlikely

LTS moment

[tardy cosmos]

21 is lts

[icy shadow]

ik but so is 17

is the point

[minor summit]

so is 8

[tardy cosmos]

tru

oops, all LTS

[atomic trail]

This is why I don't use the exceptions in there, no idea why but it just doesnt print

[minor summit]

in there you are meant to handle the exception

not rethrow it, that will cause the CF to pass it further down the chain

log it in the logger, there's probably some Logger#error variant that takes an exception

[atomic trail]

But it should still print no?

[minor summit]

no

[atomic trail]

There is, but got told to parse it in RuntimeException

[minor summit]

do not rethrow it in whenComplete

handle it

if your way of handling it is logging it, use the logger

[atomic trail]

Why would throwing it not print it though?

I thought they always did