Does the try-happens-before apply here?
#help-development
quaint basin
dry hazel
you'll need to provide more information than that
paper crest
Is acting like there are no Lombok annotations and you trying to compile without these
slender elbow
Does the try-happens-before apply here?
there is no "try" there, and the Executor interface requires a happens-before relationship
Actions in a thread prior to submitting a Runnable object to an Executor happen-before its execution begins
you are using such executor in the future callback
if you really wanna be sure just make it volatile but it really doesn't matter
thorn isle
the java memory model guarantees a happens-before relationship between any operation performed before a thread is started, and any operations in that thread; and between any operation performed in that thread, and any operation performed on any other thread after the thread terminates
executors are a higher-level thing and so the JMM doesn't make any guarantees on them, but like emily said, the executor contract itself requires them to mimic this behavior
buoyant viper
what if the memory model guaranteed java
worldly ingot
What if the memory was just the Java we made along the way?
buoyant viper
What if the memory was just the Java we made along the way?
god, all i have are horrible memories
quaint basin
GC(ZGC Major Cycles, during)
11
total
5630ms
avg time
1m1s
avg freq
Is it normal for ZGC to take this long?
quaint basin
depends on what that number is
i think i have some memory leak
My TPS was going to 5 because of the default GC
so i changed to zgc
thorn isle
major cycles generally mean that to-space was exhausted and the gc was forced to do a full stop-the-world gc that copies and defrags everything
1m is pretty high for that still especially on this small of a heap, but i could see that happening if the machine is overloaded to begin with
quaint basin
This only happens with box servers; it doesn't happen with the SMP server.
thorn isle
i don't know what a box server is
quaint basin
i don't know what a box server is
like 50-islands-server
thorn isle
do you have a spark report of it happening
neither of these reports have the numbers you mentioned
quaint basin
do you have a spark report of it happening
box2 server rn: https://spark.lucko.me/hAqyimEuLq
box1 server rn: https://spark.lucko.me/WXp6Ma1gfj
thorn isle
they still don't have the numbers you mentioned
GC(ZGC Major Cycles, during)
7
total
8740ms
avg time
2m34s
avg freq
quaint basin
they still don't have the numbers you mentioned
The tps from box2 with /spark tps had 5 just moments ago but it doesn't show up in the profiler
quaint basin
thorn isle
8 seconds for a major collection is normal
quaint basin
https://imgur.com/a/MxKEO7C
what's causing the low tps then?
thorn isle
i don't know, have you looked at the graph
quaint basin
what graph?
thorn isle
spark is a sampler
it breaks down where mspt sinks into
it really isn't very good for diagnosing gc problems
quaint basin
idk if is gc causing low tps
or if is like minecraft tick server
but smp server with the same amount of the players runs well
so i think is gc
bc smp server have more chunk loading than box servers
thorn isle
apples and oranges
two identically set up servers can perform completely differently depending on what players are doing
compare the actual mspt samples
scroll down and expand the tree
then look at what is taking time
then compare the two servers
quaint basin
two identically set up servers can perform completely differently depending on w...
But it's quite a coincidence that 2 box servers have 5 tps and the SMP server doesn't
and suppostly the box servers requires less resources
thorn isle
possibly; but if you suspect a memory leak or something gc related, you need something better than spark, especially for diagnosing zgc
zgc is weird and the numbers are difficult to make sense of since it's supposed to be completely concurrent and does weird shit like suspending high memory pressure threads
that said, especially that last bit could cause things to "magically" run slower for no apparent reason (which is why zgc is often recommended against)
i recommend you switch back to g1gc and resolve your memory issue
once it's resolved, you can switch back to zgc
the first step you'll want to do is run the spark memory allocation sampler
compare smp and the box servers
quaint basin
possibly; but if you suspect a memory leak or something gc related, you need som...
But why do you think is harder to fix memory leaks with ZGC?
thorn isle
zgc is weird and the numbers are difficult to make sense of since it's supposed ...
.
it's less transparent
it's more magical
it's more complex
quaint basin
I wouldn't know how to analyze either of them, lol
thorn isle
switch back to g1gc and run /spark heapmonitor
then keep an eye on old gen vs young gen gc's
quaint basin
Yes, I know, but I already generated the file and didn't know how to view it
thorn isle
it's an ingame command that just prints the gc events in chat
analyzing an actual gc log, while the best approach, is beyond the scope of what i'm going to handhold anyone through here
thorn isle
then keep an eye on old gen vs young gen gc's
if you see a lot of slow old gen gc's back to back, you're either allocating too fast or don't have enough free heap
increase the heap and if you just see the old gen size grow proportionately to that, it's probably a leak
if it helps, it's probably allocation pressure
for the former, take a heap dump; for the latter, take a memory sample with spark
either way, the first thing to do is to take a memory sample with spark on both servers to get a baseline of both
you can send it here and i can look at it
quaint basin
I'll leave one box server with ZGC and another with the old GC
quaint basin
switch back to g1gc and run /spark heapmonitor
spark heapsummary or heapdump?
heapmonitor doesn't exists
quaint basin
spark gcmonitor
quaint basin
something monitor
idk how can i analyze this
https://imgur.com/a/qYxZknY Did this stop the main thread for 36ms?
https://spark.lucko.me/wxbbJ3GOzI idk what should i do xd
I wanted to create more servers but there isn't enough RAM lol
green mauve
what to do when the configuration is kind of complex (entities, itemstacks, a lot of stuff) and its a list of some kind of item. but you dont want to provide any default values so its just empty [] or {}. but you have to explain the user how to use it, show an example of a configuration.
i dont think it will work just easily with comments, becausae when saving, its hard to preserve them, and if you just dump them again each time saving - the user wont be able to remove the comments once they dont want it
kind hatch
what to do when the configuration is kind of complex (entities, itemstacks, a lo...
Assuming you are talking about yml files, comments have been a solved issue for a long time.
They won't be erased/limited to the header anymore.
As for explaining to the user, this kind of depends on how your config works.
Are the values absolutely required? Is there a template file provided they can reference?
What do you have currently?
slender elbow
you're going to spend a full two hours of your life trying to write the most appropriate comments or documentation possible as to how the configuration should be written for no-one to read it because everyone has a severe case of tunnel vision
green mauve
they are preserved. but when you serialize it back to yaml. then they''re kinda not..?
green mauve
you're going to spend a full two hours of your life trying to write the most app...
but anyways there has to be an example
i cant just put
something: {}
they dont know what to put there
slender elbow
which is why you are going to spend two hours of your life writing something people are going to pretend does not exist
kind hatch
Are they editing serialized data directly?
green mauve
and if i just put comments, then deserialize the yaml file, so i can access the settings in code. then its not preserved i guess, also if they remove the comment they will get it back
green mauve
which is why you are going to spend two hours of your life writing something peo...
i dont have to spend 2 hours. i just provide an example which is as simple as creating an instance of the setting
and converting to yaml as a comment
ive done that already but the comments are either lost. or the user will be stuck with them. i guess ill go with the second one
i think rubber duck method has just worked for some reason
pseudo hazel
just add the comments each time, why would the user want to remove comments
pseudo hazel
yeah exactly
its the only way to make it work normally
you could always just put a link at the top to point to proper documentation
to avoid large comments
kind hatch
Simple projects can still be complex
pseudo hazel
and then just say, if you wanna know how to configure it, take a look at the example file
green mauve
*Simple projects can still be complex*
no time for that
kind hatch
I'm wondering if there's a better way to go about configuring whatever it is they are doing
green mauve
what do you mean?
pseudo hazel
if its in-game stuff like item stacks I usually have a way to set those things in-game
green mauve
im making an abstract YamlConfiguration class
green mauve
if its in-game stuff like item stacks I usually have a way to set those things i...
yeah
i dont think its matters for the topic i mentioned though
oh
it does
but not everything is that easy to do in-game
kind hatch
Well, like Steaf said, in game configuration will beat out file based configuration (for the average user at least), but when it comes to files, keeping them simple still goes a long way.
green mauve
yep. but
kind hatch
It sounds like you are making them edit serialized data directly, but that's my assumption.
I don't know without an example file
green mauve
i actually cant think of a clean way to do things like these
- region:
world: world
region-name: region1
mobs:
- drops:
- type: DIAMOND
amount: 1
meta: {}
- type: GOLD_INGOT
amount: 1
meta: {}
- type: IRON_INGOT
amount: 1
meta: {}
type: ZOMBIE
and its not most complex
but still its easier to do with file
pseudo hazel
yeah so its a custom itemstack serializer I guess
green mauve
ye
pseudo hazel
/plugin mobs ZOMBIE drops set
and then take items from the players hotbar or whatever
but yeah it still depends on your goal
green mauve
then you add chances, damage, health. i dont think it would be as easy for the user and me to put it in-game than just going for the file, it becomes a tree
pseudo hazel
true, file will be easier to create as a dev
green mauve
allright
kind hatch
Yea, I guess if this is targeted more so at devs/administrators, then there is the assumption that they should know enough to work with this format.
pseudo hazel
that too, but I was also more just talking about the ease of coding it
mortal hare
yeah so its a custom itemstack serializer I guess
isnt there a standardized serializer for this like
green mauve
yep, i just dump a Settings object or some MobData object into adpater and yeah thats it
pseudo hazel
since you said you have limited time
mortal hare
we solved chat formatting inconsistencies with adventure api and minimessage
pseudo hazel
isnt there a standardized serializer for this like
maybe but iirc its not perfect
mortal hare
so why isnt there a standard format for serialization of items
green mauve
dunnoo
pseudo hazel
because items are annoyingly complicated
kind hatch
Cause it's hard
green mauve
yeah the nbt or smth
pseudo hazel
serializing to bytes is still the best way to store items xD
green mauve
and most readable
pseudo hazel
also for the dfu and stuff
green mauve
🐈
kind hatch
Plus any changes to data structure
pseudo hazel
well a human reading bytes is like the least readable
mortal hare
actually byte serialization that java has introduces some rce vulnerabilities if those configs get out of the server
since java's serialization of classes is kinda weird and doenst protect from custom data being injected, so you could theoretically load a class instance which wasnt even supposed to be loaded in a first place
green mauve
well a human reading bytes is like the least readable
i like to read bytes
they have a deep meaning
slender elbow
so why isnt there a standard format for serialization of items
there is it's called nbt and snbt
slender elbow
how is snbt any different than yml
it's perfectly fine
nbt is data because it's data, yes
echo basalt
so why isnt there a standard format for serialization of items
I kinda wrote one
well I wrote an api that looks a bit like nms' and can decode items and components from nbt, yaml and json
ivory sleet
they suck in general purpose use
how?
thorn isle
strictyaml -> json -> json nbt ops -> nbt
young knoll
Ah yes more conversion layers
thorn isle
my_helmet:
components:
minecraft:trim:
pattern: "spire"
material: "emerald"
minecraft:dyed_color: 0x636C6D
minecraft:max_damage: 550
yes, all threads
did you take a memory sample yet?
echo basalt
```yaml
my_helmet:
components:
minecraft:trim:
pattern: "spire"
...
p much
buuut to make it nice for users let em have aliases and stuff
also where's the fun in not having custom components
:)
umbral ridge
```yaml
my_helmet:
components:
minecraft:trim:
pattern: "spire"
...
thats pretty cool
quaint basin
did you take a memory sample yet?
not yet
see this: https://imgur.com/a/IqTcbzg
CPU usage went up to 80% when TPS went up to 15, probably to clear memory
quaint basin
see this: https://imgur.com/a/IqTcbzg
CPU usage went up to 80% when TPS went up...
this is from spawn server, meaning it doesn't seem to be a memory leak from the box servers; or perhaps the CPU usage of the box processes caused the TPS to drop on the spawn server
thorn isle
the gc going crazy on a server because of a leak/too much pressure will cause cpu usage to skyrocket
so that could still be the cause
but yes, the box being overloaded does reduce tps
quaint basin
I'm getting UserNotFoundException on player join.
How is that possible?
i just remove on playerquit
I thought I had guaranteed visibility on playerjoin
grim hound
how do I enable debug?
grim hound
https://pastes.dev/SjKQ5MKdi4
I'm getting UserNotFoundException on player join.
cuz
playerjoinevent
does not need to be the first event
what's more
it isn't
bitter hazel
Sorry if wrong channel, looking for Minecraft devs for VNX SMP (AH, Economy, GUI,Commands) if you are interested DM me
chrome beacon
?services
undone axleBOT
If you wish to request or offer development/art/building/administration services, please do so at https://www.spigotmc.org/forums/services-recruitment-v2.54/
quaint basin
I think you misunderstood. The playerjoinevent is not reading the map#put made in playerasyncprelogin.
thorn isle
let's see your asyncprelogin
slender elbow
I'll show you mine if you show me yours
thorn isle
what's the name of that swelling up blushing emoji again
chrome beacon
thorn isle
mm yes
quaint basin
let's see your asyncprelogin
i dont send the event listeners for some reason im dumb
quaint basin
https://pastes.dev/SjKQ5MKdi4
I'm getting UserNotFoundException on player join.
I just have userService#remove on PlayerQuitEvent
thorn isle
when in doubt, print sysouts
you're sure the uuid/name isn't being changed after the listener inserts them into the map?
quaint basin
you're sure the uuid/name isn't being changed after the listener inserts them in...
The name doesn't change anything on this error
Is it possible for the UUID to change between asyncplayerpreloginevent and playerjoinevent?
Isn't he always the same?
thorn isle
the uuid in the prelogin event can be changed
thorn isle
by someone changing it, for whatever reason you'd change it, e.g. mixed online/offline mode or username changes
quaint basin
🤡
My 99 plugins are so poorly made then
Which event allows me to change the UUID?
I can't do event#setuniqueId in asyncplayerpreloginevent
thorn isle
print sysouts
rough sonnet
you can change the profile iirc, that includes the name and id
quaint basin
So what should I do to avoid this then?
mortal vortex
I can't do event#setuniqueId in asyncplayerpreloginevent
Are you on Spigot?
chrome beacon
They're not
quaint basin
Are you on Spigot?
yes
mortal vortex
Paper has a HandshakeEvent which might be messing with you
buoyant viper
when in doubt, print sysouts
truth nuke 🙏
quaint basin
on really im running paper yes
mortal vortex
Dont say spigot then cuh
buoyant viper
and then paper yells at u with authornagexception for sysout
mortal vortex
fuck paper
quaint basin
maybe i'll need get/create data from db on playerjoin then
and prohibit interactions with the plugin until the database fetch is successful
mortal vortex
I think on paper you can do:
@EventHandler
public void onHandshake(PlayerHandshakeEvent event) {
event.setUniqueId(idk);
event.setPropertiesJson(event.getOriginalPropertiesJson())
}
is the handshake id respected? isn't that just the crap the client sends
which is ignored because the client is authed anyways
not sure if paper decided to ignore the protocol and have their event work entirely different
yeah there's setPlayerProfile in the AsyncPlayerPreLoginEvent
buoyant viper
fuck paper
true... Spork would never nag u for using sysout
mortal vortex
FUCKING TRUE
buoyant viper
thank you paper, very descriptive
almost all the javadocs in bukkit api are this also lol
rough sonnet
would be nice to know if the uuid is actually used anywhere, but regardless
the async prelogin event has a profile setter
slender elbow
yeah i mean this is one of those cases where you'd use monitor priority
buoyant viper
thought i was gonna be able to solve a design problem im facing by realizing i used a Map that where the Key is the information i need for how im using the Value
little did i know... past me hid the map behind implementation and only expose a Collection of Values as API
and i kind of dont want to change that
but if i dont, then im holding onto a reference to an object for waayyyy longer than i actually need to
buoyant viper
does a plugin that depends on another plugin not only wait for it to be loaded, but also enabled?
lilac dagger
It should yes
Worth to test tho
I think that plugin loads yours load that plugin enables, yours enables
buoyant viper
i hope so
lilac dagger
I'm pretty sure it'd be reported as a bug if otherwise
You could make 2 plugins with names annabelle and xenia
And then depend on each
Actually just xenia
And send a log message on load and enable
But not entirely sure if the plugins are sorted by names
buoyant viper
And then depend on each
cyclic dependencies :(
lilac dagger
No
I meant individually
And also it's not cyclic
Annabelle will run first
Then xenia
And that yours
They're is no annabelle depend on yours
So that there's a cyclic dependency
buoyant viper
if they depended on each other that would be cyclic
buoyant viper
hmm
lilac dagger
If your is named b to y
And the dependencies are named sorted
Then this'd be an easy way to test
wet breach
pretty sure dependencies are loaded in the order they are defined
whether it be in the load thing in config or the manifest file
lilac dagger
Ah no
There's no config or manifest, dependencies (plugins) are dynamic
So a file wouldn't work for this
You can look in simple plugin manager
If someone has access
wet breach
are you talking about the order of the plugin list?
if so, order of the plugins loading is not guaranteed, however if nothing is defining a dependency to load in the plugins being loaded, then it typically defaults to being alphabetical since that is the general standard of how directories are listed
IE, its mostly OS dependent how the directory is listed and the JVM doesn't really alter that unless told to
buoyant viper
pretty sure dependencies are loaded in the order they are defined
does this also carry through to enable order?
ie. if A depends on B, would B#onEnable call before A#onEnable?
?stash
undone axleBOT
wet breach
Depends what kind of dependency we talking but in general no. If you don't have it defined you are depending on another plugin, your plugin will be loaded and then ran hence why it has always been a thing to structure your plugin to handle waiting on dependencies to be loaded or to not run anything that depends on something within the first few minutes lol
because there is no guarantee that right after yours is loaded that your dependenices loaded before hand, or will load immediately after
buoyant viper
Depends what kind of dependency we talking but in general no. If you don't have ...
plugin with a hard-depend on another plugin
wet breach
if its defined in the config, then the server will not attempt to run your plugin until what you defined has been loaded as far as I am aware
that is the only thing that alters the loading order from plugins
otherwise its whatever order the JVM obtained from the directory lol
buoyant viper
good enough for me
wet breach
you can perform a test if you want, but just know it would be platform dependent as well because as I said, the directory lististing is depending on how the OS responds to the request and how it chooses to order everything. From windows the default comes from FindFirstFile function
which is the function that controls directory listing in windows
buoyant viper
ye ik that dir listing isnt guaranteed to be sorted
i just didnt know if plugins wait for their dependencies to be enabled before they enable
wet breach
nope, not unless the server is told to do so
buoyant viper
or if there was a chance a dependency could be loaded, but not enabled, before the dependant enables
wet breach
otherwise it will attempt to load yours and enable, and then it will fail if you required something and then move on, which if you are not careful you can cause a cascading effect if you have enough plugins depending on each other
IE, one plugin failed to enable, because something was not enabled yet because your onEnable code required it, and then this causes another plugin to not enable because this one didn't etc
buoyant viper
update i ended up with something similar to this (but 100x more autistic) bc i realized theres a chance rounding errors with valid doubles but invalid "units" (stuff like .001 and .009 had odd rounding, and even 0.01 acted strange)
so now im going string <-> long instead of string <-> double <-> long
(i thought i was safe doing * or / 100.00, i was not)
wet breach
lol
buoyant viper
oh god it shit the bed
What
how the fuck
surely it cant be my '\0'...
god damn it it is
theres rly no way to have replace(char, char) just nuke a char from existence? damn
whole Strings it is i guess
pseudo hazel
can you split on the char and then join it again? xD
buoyant viper
can you split on the char and then join it again? xD
fuck ur so right
u can only split by whole strings 😭
ill just have to write my own .replace method...
buoyant viper
it appears in my hubris i didnt think "" was some fuckshit in java
(it is zero-length whereas '\0' still counts towards the overall string)
wet breach
it appears in my hubris i didnt think `""` was some fuckshit in java
this is the equivalent of making an empty string
buoyant viper
yeaaahh
actually wait it wasnt me misunderstanding "" was '\0' that i was using the wrong way
ofc "" makes sense bc then im actually telling java "replace this with NOTHING", whereas '\0' IS still something
and for some reason i thought that would be equivalent to using an empty string
?jd-s
undone axleBOT
buoyant viper
ill just have to write my own .replace method...
i wonder how fucky i could make this
im imagining some dumb shit specifically for null char like
String replace(String str, char oldChar, char newChar) {
char[] out = new char[str.length()]; // worst-case where len(out) = len(in)
int len = 0;
for (char ch : str.toCharArray()) {
if (ch != oldChar) {
out[len++] = ch;
} else if (newChar != '\0') {
out[len++] = newChar;
}
}
return new String(out, 0, len);
}```like 3 array copies tho kinda pepega (out alloc, toCharArray, and new)
buoyant viper
john java?
call me john java the way i java'd that john
if java had mutable strings that could be single alloc but noOoOo
wet breach
you can however use binary for this
lilac dagger
im imagining some dumb shit specifically for null char like
```java
String repla...
pretty sure string builder has remove by index
buoyant viper
oh wait StringBuilder will also just make its initial capacity the string length (+ 16 for encoding i guess?)
oh but deleteCharAt does some weird arraycopy stuff to shift everything
buoyant viper
i might have best-case here somehow lol
lilac dagger
it looks fine
just that instead of doing manual replace
you can do removeFromString(String string, char character)
you don't need to do it for \0
buoyant viper
so i did consider that
and is probably what i would end up using
since duplicate replace method makes no sense
lilac dagger
you only need to if (ch != character)
buoyant viper
ye
lilac dagger
and then increment and add current char to your array
buoyant viper
yeah :p
again, java, id kill to be able to work on the backing array for strings... just this once
lilac dagger
it also handles utf 16
buoyant viper
its funny bc i think String#replace(char, char) itself doesnt care about encoding
oh nvm it does
lilac dagger
it does yeah
quaint basin
thank you paper, very descriptive
The player's UUID after the player join event won't change until the player quit event be called, right? I know the question is stupid, but I don't doubt anything anymore.
ivory sleet
StringBuilder also is full of reallocs :(
thats usually fine
since it uses system.arraycopy
which is quite fast
(obv in terms of pure theoretical time complexity its less good, but in practice it tends to work out just fine)
slender elbow
it also does not realloc if it doesn't need to
deleteCharAt does not, for example
it doesn't create a new array or anything
hazy parrot
The player's UUID after the player join event won't change until the player quit...
Player UUID is never changing
echo basalt
any of y'all know any tools where I can toss code with spigot mappings and it auto converts
thorn isle
The player's UUID after the player join event won't change until the player quit...
Are you printing sysouts yet
buoyant viper
deleteCharAt does not, for example
agh
true
i didnt notice the arraycopy was to the same array
(idk where i thought an extra array was coming from)
maybe this Isnt something i should worry about also
its not like these strings are gonna have input values that are like 1 Billion characters long
i should Not waste my time trying to find an optimal way to replace one character in string whos length will probably never be over 10
young knoll
Clearly needs JNI and hand optimized assembly
buoyant viper
Clearly needs JNI and hand optimized assembly
TRUE!
im actually creating a native binding as we speak
buoyant viper
following stdlib callchains is... something
we got the StringBuilder(String) { AbstractStringBuilder(String) { this.value = new byte[String#length() + 16], AbstractStringBuilder#append(String) { String#getBytes(this.value, PARAMS) (redirects to [StringLatin1|StringUTF16]#getBytes) { for (;;) { dst[i] = src[i] } } } }
manual copy in append but arraycopy in deleteCharAt
(most likely due to how a string encoder gets a byte at index)
it all makes sense but my lord
i think i only have one option @lilac dagger @young knoll ... i Need to fork OpenJDK and submit a PR
we need String#remove(char)
(it will just call String#replace(valueOf(char[] {char}), "")
glad berry
buoyant viper
and String#split(char, int) (it just calls String#split(valueOf(char[]{char}), int)) + overloads
buoyant viper
crazy revelation
String#charAt is copyless
String remove(String str, char ch) {
int max = str.length();
char[] chars = new char[max];
int len = 0;
for(int i = 0; < max; i++) {
char c = str.charAt(i);
if (ch != c) {
out[len++] = c;
}
}
return new String(chars, 0, len);
}```this is Great news @lilac dagger
(there is no circumstance where i should actually use this method)
thorn isle
charIterator might inline better
buoyant viper
better idea
String.class.getDeclaredField("value").set(str, null);
Nuke the string
curse u java 9+ modules, i cannot access java.base
ivory sleet
worse comes to worst, unsafe might do the trick :>
not sure if strong encapsulation protects against it
buoyant viper
I CLOSED INTELLIJ AND I GOT 30 GB OF DISK SPACE BACK??? AND 20 GB OF RAM????
JETBRAINS
WHAT?
young knoll
You never just make 30 GB of lock files?
buoyant viper
😭
lilac dagger
Needs more ram in this economy?
buoyant viper
https://github.com/Fr33styler/BenchmarkZone/blob/main/src%2Fmain%2Fjava%2Fro%2Ff...
im stealing that thx
lilac dagger
Not sure it's gonna be useful to you tho
thorn isle
intellij please
this is also gradle running out of heap, not ij
gradle likes to eat all available compute memory and disk space
because that's what build tools are supposed to do, right
slender elbow
i really wonder how you get to do that lol
I keep using a 8gb machine for Android work and it works fine
echo basalt
most I've pulled from IJ was 50gb
cloud perch
most I've pulled from IJ was 50gb
slender elbow
I always keep it on the default limit from toolbox and then Xmx2G for gradle
cloud perch
how much ram do u have
echo basalt
64gb on my desktop rn
slender elbow
i think on my PC toolbox's default id 2gb, and on work PC it's slightly less than 1gb because it has mercy on me
echo basalt
laptop has 16gb soldered 😢
slender elbow
so I just leave it at that and it just works lol
cloud perch
64gb on my desktop rn
echo basalt
I've been also messing around with running my own little copilot thing which pulls 11gb vram
cloud perch
ddr4?
echo basalt
yeah I built this a while ago
cloud perch
ahh thats why
echo basalt
I'd go for 96gb ddr5 if I were building this today
cloud perch
the ram prices are still high 🥀
cloud perch
I mean thats pretty fair
echo basalt
then they became like 30 bucks
cloud perch
💀
slender elbow
just build your own ram
echo basalt
I was debating running 256gb ram for shits n giggles it was like 150 bucks
slender elbow
robust helm
just build your own ram
hi fellow hackernews reader
robust helm
huh??
robust helm
not even in my history lol
then its a crazy coincidence cause check the #1 post on the front page today
its exactly that video
slender elbow
okay
lilac dagger
just build your own ram
I am yet to build my own 1 bit nand
Imagine building 16 gigs of it 💀
robust helm
just use swap 😛
robust helm
degrading ssd?
slender elbow
that's why i use a hdd for swap
lilac dagger
Imagine building 16 gigs of it 💀
But once you build 1 gb of ram i guess the mechanism is just to add one more block
Probably even 16kb blocks is enough to repeat
robust helm
degrading ssd?
what does it mean?
lilac dagger
that's why i use a hdd for swap
Sadly my laptop doesn't come with a hdd
chrome beacon
what does it mean?
SSDs wear out when writing to them
lilac dagger
what does it mean?
Ssds have limited write before they're burned
Same mechanism as batteries i think
robust helm
SSDs wear out when writing to them
oh fr? i thought thats just a hdd thing
slender elbow
run smartctl for more info !!
robust helm
So anode and catode gets oxidized
uhh yeah whatever you say
chrome beacon
oh fr? i thought thats just a hdd thing
HDDs don't really have that issue (at least not as much, they last a lot longer). Looks like you got them confused
lilac dagger
Well there's videos where you see it in action with batteries
robust helm
HDDs don't really have that issue (at least not as much, they last a lot longer)...
yeah probably
chrome beacon
I've been using my boot ssd for 13 years 💀
lilac dagger
It's probably a slc
Or wait
What's the term when a trap only holds 1 bit?
Oh yeah
Slc
Modern ones most likely are mlc
They wear faster but are cheaper
And also nobody bothers to put this technology on the store page when you buy an ssd
chrome beacon
53% health 😭
slender elbow
53% health 😭
that's fine you still got more than half of its lifetime yet to use lol
lilac dagger
But i don't know what's this endurance about
chrome beacon
It's just what it's rated for
chrome beacon
Yeah it's not like the drive will explode when I reach it
Still though I really should swap
if only the ssd prices weren't so bad 😭
lilac dagger
I'd still try to keep backup data
chrome beacon
I only really have Windows on there
lilac dagger
Oh it's fine
chrome beacon
It's 128gb doesn't fit much else
lilac dagger
Ye
Same for me
If you have an hdd also you could move the swap space there
I did it for both windows and linux
thorn isle
i recommend against it, hdd is too slow for swap to be very useful, it will take a long time for paged applications to "wake up"
it is better than crashing from out of os memory, but only marginally
thorn isle
something that works (and which we used back in the day) is to plug in a flash drive in a usb 3 port
lilac dagger
Is that fast enough?
thorn isle
much faster than a hdd
lilac dagger
Nice
thorn isle
a flash stick is essentially a small cheap ssd with fairly limited bandwidth
lilac dagger
But i can't do it on a laptop
thorn isle
does your laptop not have usb ports?
lilac dagger
It has but it's annoying to have ports with things on them for long
Plus i move it around i don't wanna damage the port
I have to usb 3
Both are blue
Wait
I also have a microsd slot
Can't i put one there?
thorn isle
check the specs and compare to the hdd
lilac dagger
Theoretical max speed for micro sd is 300
If using the best one
I assume it's probably around 200
MB/s
A western digital one is saying speed of 173
thorn isle
a big issue with hdd's and optical disks, both being things with moving parts, is seek times
they perform relatively well for contiguous reads/writes, but random access involves a lot of physical movement and seeking
which is bad for random access memory, i.e. ram, in particular
thorn isle
not totally sure how it works in conjunction with paging; bringing the page back into physical ram is a contiguous read i'd assume
lilac dagger
Takes time
Yeah
But also a nice thing about hdd is the write and read is quite similar no?
pine crest
not totally sure how it works in conjunction with paging; bringing the page back...
ideally, but doesnt have to be
lilac dagger
Flipping a magnetic bit is fast i assume
Windows does and linux does allocate space
So i assume it's all in one place
Unless you allocate space on multiple drives (i assume it's possible)
ivory sleet
So i assume it's all in one place
i think what cakeeater meant to say is that it might be split behind several pages right?
lilac dagger
Too many assumptions
ivory sleet
ofc depending on page size ^^
lilac dagger
I don't know what a page is
ivory sleet
memory pages
lilac dagger
Is that a ram memory block?
thorn isle
roughly yes
ivory sleet
kinda
lilac dagger
Oh i see
ivory sleet
i think the primary way it makes sense is that a pages allow contiguous virtual memory to not be contiguous physically
thorn isle
looking at some forums, it looks like people advise against sd cards in particular because they don't disperse/manage writes as well as proper ssd drives, and so wear out much faster from the constant writes when used as swap
lilac dagger
looking at some forums, it looks like people advise against sd cards in particul...
Good to know
thorn isle
nobody seems to have anything to say about the performance, everyone is being a reddit nanny and going "nooo you can't do that you will wear out the card"
but from the numbers i did see it probably does perform better
lilac dagger
i think the primary way it makes sense is that a pages allow contiguous virtual ...
As in ram you mean by physically?
thorn isle
i fucking hate people that start nannying instead of answering the asked question
ivory sleet
As in ram you mean by physically?
well you're aware of how physical and virtual adresses work right?
lilac dagger
So you can have partial reads from swap instead of the whole page?
lilac dagger
well you're aware of how physical and virtual adresses work right?
No, teach me
ivory sleet
u have the virtual address space, that is the addresses we write code against, then those addresses are translated to physical addresses
lilac dagger
I know what adresses are and rough details about ram
ivory sleet
this is for address protection, but also to allow for example paging
lilac dagger
So abstracting
umbral ridge
sd cards arent designed for constant writes.. well.. depends which ones
ivory sleet
yea
lilac dagger
In a sense
ivory sleet
but RAM strictly speaking is physical memory
lilac dagger
I see
umbral ridge
they wear out faster than ssds
lilac dagger
Yeah makes sense
lilac dagger
they wear out faster than ssds
But how big is the nand on a stick?
I guess i have to open up one to see
But yeah if you think about it a microsd doesn't have a lot of real estate
And they also boast huge memory sizes
lilac dagger
but RAM strictly speaking is physical memory
Thanks for teaching me about virtual memory
umbral ridge
usually usb has a smaller controller chip .. maybe 1 or 2 nand flash packages
ivory sleet
Thanks for teaching me about virtual memory
haha yea, and so the idea is that u have these pages, so basically ur app (which might live between 0x00000000 and 0x0000FFFF) looks to be contiguous, but physically, it might not be at all
lilac dagger
But ram is fast
ivory sleet
yep
thorn isle
basically each application gets its own virtual memory space that starts from zero, and then that gets mapped, in increments of pages, to physical memory
e.g. the way how java guards against null pointer references is by having the null pointer point at memory address zero, which is in page zero, which is set to be inaccessible; so trying to access it raises an error which java handles and converts into a NPE
ivory sleet
^ and when u for example swap processes in and out, there's a chance once a process is reloaded, it will not be contiguous physically post-swap, but ofc virtually it will be.
lilac dagger
basically each application gets its own virtual memory space that starts from ze...
Pretty cool
thorn isle
this was originally just for managing the virtual memory mappings, but eventually someone realized that you could shove infrequently-used pages to disk into a file
ivory sleet
^
thorn isle
and then windows called it the pagefile.sys
lilac dagger
😄
thorn isle
but since hdd's which were essentially the only thing that existed at the time were slow, it became common practice to use external flash drives to store the pages on
ivory sleet
paging is used to solve if u run out of memory (where it will trigger a bunch of page faults, basically swapping pages in and out frequently), but also to minimize unused memory whilst allowing for locality
thorn isle
which is why if you plug a flash stick into windows xp, it asks if you want to "use this drive to speed up the computer"
slender elbow
nobody seems to have anything to say about the performance, everyone is being a ...
what's the point of having one if you can't use it !!
i will use the sd card until it melts
slender elbow
which is why if you plug a flash stick into windows xp, it asks if you want to "...
oh my god memory unlocked
lilac dagger
Pretty cool stuff
lilac dagger
usually usb has a smaller controller chip .. maybe 1 or 2 nand flash packages
Still, even if it's jusr one it's bigger than a micro sd card
umbral ridge
why not just use an ssd
lilac dagger
Well all this started because of a ~50% durability left on a ssd with windows
And endurance at its limit
So we talk about alternatives and ended up learning about virtual memory
umbral ridge
that must be a really old ssd
lilac dagger
And how physical memory is not guaranteed to be continuous
10 years or so i believe olivo said
13
umbral ridge
yea 50% is good for 13 years..
although, i have m.2 ssds from samsung from 2019 and they're all still on 95% and above
kinda depends what u use it for i guess
i dont think theres anything that would work permanently and not have a durability issue
everything wears as time passes
lilac dagger
Yeah
If you use swap space on it
And get freqently over your designed ram capacity
Is one way to lower it i think?
lilac dagger
i dont think theres anything that would work permanently and not have a durabili...
Yeah
I don't know how the trap on a ssd work
But pretty sure it degrades in time
thorn isle
having the ssd full near its storage capacity also accelerates how fast it wears
lilac dagger
Batteries do
lilac dagger
having the ssd full near its storage capacity also accelerates how fast it wears
Interesting
thorn isle
i find it kind of funny how the solid state thing wears out faster than the thing with the moving parts
lilac dagger
Well, if smarter people than me solves batteries pretty sure we'd see this in ssds too
If it can get this small that is
I'd love to have a phone for 50 years and still have a charge
Not talking about free energy
Just to keep it in a shelf and be usable in 50 years
But sounds impossible
umbral ridge
yeah its more about chemistry than size.. batteries just degrade over time no matter what.. even if you dont use them, they slowly lose capacity
batteries kinda destroy themselves over time
lilac dagger
Yeah sadly
robust helm
batteries kinda destroy themselves over time
this sounds so sad
buoyant viper
i really wonder how you get to do that lol
starting and stopping a debug minecraft client environment multiple times in a row over a long session
slender elbow
yeah I mean been there done that, but haven't had issues like that lol
buoyant viper
and honestly neither have i
but newer intellij has been shitting the bed on me a lot (even if this is gradles fault)
buoyant viper
yeah I mean been there done that, but haven't had issues like that lol
it might be a bug the specific gradle version im on, bc this was a non-issue before i decided to update my wrapper
bc this project is nearly 8 years old
(i went from gradle 6.7 to 6.9.4)
(its java 8)
(its minecraft 1.8.8)
buoyant viper
?jd-s
undone axleBOT
young knoll
It’ll throw an exception
quaint basin
Chunk system has shut down, cannot process chunk requests in world
'5fb6d722-8f05-43f0-be8c-1dba06d8fbcc' at (0,0) status: minecraft:full
What does this mean? I got this on playerquitevent
buoyant viper
It’ll throw an exception
what the fuck why 😭
young knoll
Idk
Maybe the items aren’t in the world yet
And spigot didn’t really have support for that back in the day
slender elbow
i mean you can just call world.spawnItem or whatever
iirc that list is actually backed by Item entities, but only the ItemStacks are made available, it's a lazily transformed list so you can't really add to it
buoyant viper
ah
yeah if it was just itemstacks it made like no sense to me as to why it was only partially mutable
wet breach
Just to keep it in a shelf and be usable in 50 years
this is doable but you need a special battery where shaking it will give you a charge
they exist if you can find them
lilac dagger
Sort of like a bike's 'alternator'?
But i think all batteries get oxidized with time
You still need to store that energy that you get while shaking
But it'd definitely work
wet breach
You still need to store that energy that you get while shaking
well its chemical reaction
you need a chemical that will react to a substrate, chemical can sit at the bottom of battery
shake it, and due to the shaking and interaction with substrate you now have a charge
thorn isle
we've had self-winding watches for a long time
slender elbow
I like the convenience of smartwatches but mechanical watches are just sexy
sly topaz
is it even convenient
only time I find it useful is for tracking your sleep or health stats when you work out
analog watches are completely useless to me, I can more or less tell the time by looking at the sun
eternal night
sly topaz
they are definitely a good accessory, specially for men
I don't see women wearing them all that often, even though there are some really delicate ones that could really elevate the elegancy of your outfit
slender elbow
the things I find convenient are sleep tracking, alarm clock (bonus for it being based on your sleep state before alarm time) (i find them vibrating on my wrist much more effective than a regular alarm clock), and being able to receive notifications directly on the watch without having to look at my phone and assess if I want to bother with it (bonus points for being able to filter per application or contact)
i don't really use it for anything else besides timer/stopwatch when I don't have my phone around
oh and of course find my phone if you are very silly
sly topaz
I got a apple watch se 2 sitting on a drawer
I used to use it for sleep tracking but I managed to fix it without looking at the numbers so I am not too interested in it anymore
though I haven't tried anything fancier so maybe I am biased
lilac dagger
I like the convenience of smartwatches but mechanical watches are just sexy
I like them for the mechanism
The one with the heart that spins
Forgot how it was called
That makes sure it doesn't lose precision as much due to gravity
slender elbow
though I haven't tried anything fancier so maybe I am biased
oh i have a fitbit versa, this is far from fancy
thorn isle
i use a 5 dollar digital casio watch that i originally got for farm work because my actual mechanical watch would be too expensive to be regularly caked in shit and grime and banged into things
however i stuck to it because it has an egg timer in it
slender elbow
😌
quaint basin
Will the chat event be removed?
chrome beacon
Read the javadoc
quaint basin
Read the javadoc
i cant find any info about that
chrome beacon
if it doesn't say then no
buoyant viper
*laughs in "think of the children" laws*
fuck them kids..
for legal reasons, i do NOT mean "fuck" as in "to have sex with". i am NOT advocating to commit illegal activity with minors
thorn isle
it's deprecated because it's superseded by the async chat event and eventually the 5 other adventure chat events of various synchronicities that paper adds twice per annum
it works but it's not ideal as it causes an unnecessary task to be scheduled on the main thread
worldly ingot
And also halts the server thread if you're doing something stupid in your listener
Basically, use AsyncPlayerChatEvent (and recognize you're in an async context!), or if you're using downstream software, use one of the dozen chat events they have for some reason
buoyant viper
im gonna make an AsyncPlayerChatEvent listener that just calls Thread.sleep to make the player think they're lagging
buoyant viper
i want to lag the player not the server :(
lilac dagger
oh alr
slender elbow
I mean there are still reasons to use the sync chat event, but those are the exception and you should default to the async one
the event still exists for those valid reasons
lilac dagger
i just use the async event and call a runtask for the next tick if it's the case
slender elbow
yeah but that's cringe if you need the sync result to mutate the event
lilac dagger
ah no
it's just a one way out to things like unsafe collection for example
hmm
if i think about it, i should maybe use the playerchat event
quaint basin
why would it be removed
i mean sync event
quaint basin
it works but it's not ideal as it causes an unnecessary task to be scheduled on ...
I ask this because my task needs to be done in the main thread anyways
thorn isle
does it need to mutate the event?
quaint basin
does it need to mutate the event?
wdym
thorn isle
you're writing an event listener
quaint basin
no
thorn isle
why are we talking about events
quaint basin
I have a method that I need to call in the chat event, and if I called it using asyncchatevent, I would have to schedule a task to be executed in the main thread; but then I should use chatevent
My fear was that the event would be removed since it's deprecated
thorn isle
does the task you're firing from your chat event listener need to be able to modify the state or results of that event?
quaint basin
No but my method is not thread-safe
thorn isle
then use the async chat event and schedule a task
quaint basin
then use the async chat event and schedule a task
why not use the chat event sync?
thorn isle
because the server will block for it to complete so that you can modify the event
which is unnecessary
quaint basin
i dont understand
thorn isle
chat is async
regardless of what event you listen to
so if you do listen to the sync chat event, the server has to schedule that to run on the main thread and wait for you to be done with it before it can e,g, send the chat message to players
it does this because event listeners are expected to be able to modify the event
however your task does not need to modify the event
so this is redundant and unnecessary
thorn isle
what is blocked?
quaint basin
idk the main thread
thorn isle
no
your task runs on the main thread
the chat logic can complete freely and regardless of what you do on the main thread
quaint basin
I really don't understand the difference between listening to a sync event or listening to an async chat event and scheduling a task there
thorn isle
one couples the lifecycle of your task to the chat event and the other does not
the chat event cannot complete until your task is done if you just run it inside the sync listener
quaint basin
the chat event cannot complete until your task is done if you just run it inside...
And until the chat event is complete, can the messages not be processed by the "async" threads?
thorn isle
the sync or the async chat event
thorn isle
okay so let me put it like this
you have a fridge at home
you go to your fridge to see if it has cheese in it
you find cheese in your fridge, and now intend to eat it
do you
a) take the cheese and eat it
b) go to the supermarket with the cheese, eat the cheese there, and then come back home
quaint basin
do you
a) take the cheese and eat it
b) go to the supermarket with the cheese, e...
a)
thorn isle
since your task doesn't need to do any work on the event, it is completely pointless to have the event wait for it, and to incur the scheduling overheads involved in having the naturally-asynchronous chat logic synchronize to the main thread
so what you want to do is to listen to the event in its natural thread, and then schedule your task onto the main thread, letting the chat event proceed without pointlessly shuffling the whole thing to MT and back
quaint basin
Are the threads responsible for processing the messages blocked until my sync chat event is complete?
thorn isle
it doesn't matter
the specific thread carrying the specific event is
that is the nature of events
quaint basin
the specific thread carrying the specific event is
In that case, isn't it always the main thread?
thorn isle
whether or not that impacts other chat events is not defined in the api; it could or it could not, iirc there is a pool or something
chat runs on its own threads
thorn isle
the difference is that you're not synchronizing the ENTIRE event to main
only your task
only the part that needs to be on main
quaint basin
the difference is that you're not synchronizing the ENTIRE event to main
Okay, that makes sense now
but
If I use asyncplayerchatevent, it forces me to use thread-safe Map structures like concurrenthashmap; if I use sync event, I only use a hashmap; perhaps the overhead of using concurrenthashmap is heavier than synchronizing the entire event
young knoll
No
thorn isle
it doesn't force you to do anything
just put the not thread safe shit in a task and run the task on main
quaint basin
just put the not thread safe shit in a task and run the task on main
this is literally the first line of the chat event block
thorn isle
the difference is, again, that it waits for the task to complete before continuing with the event execution
quaint basin
boxManager#addFriend need be executed on the main thread
thorn isle
i.e. you are delaying all chat by however long it takes for the main thread to pick up the task
quaint basin
actually cache is a hashmap bc im using sync chat event
quaint basin
https://pastes.dev/dGY7Af0Y5f
boxManager#addFriend need be executed on the mai...
if i use asynchatevent i'll need to use concurrenthashmap to cache var
sly topaz
I think I am experiencing a dejavu
young knoll
Tbf im curious why boxManager.addFriend needs to be on the main thread
thorn isle
@EventHandler
public void onAsyncChat(AsncChatEvent event) {
Bukkit.getScheduler().runTask() -> {
Player player = event.getPlayer();
UUID uuid = player.getUniqueId();
if (cache.contains(uuid)) {
event.setCancelled(true);
cache.remove(uuid);
Component message = event.message();
String target = PlainTextComponentSerializer.plainText().serialize(message).trim().split(" ")[0];
Box box = boxService.get(player.getUniqueId());
if (box == null) {
player.sendMessage(MiniMessage.miniMessage().deserialize(lang.getString("wait", "<red>Wait...")));
return;
}
SlimeWorldInstance slimeWorldInstance = BoxPlugin.advancedSlimePaperAPI.getLoadedWorld(player.getUniqueId().toString());
if (slimeWorldInstance == null) {
player.sendMessage(MiniMessage.miniMessage().deserialize(lang.getString("congrulations", "<red>You did the impossible! Please contact an administrator to report this error.")));
return;
}
boxManager.addFriend(player, target, box);
}
}});
What is it doing that isn’t threadsafe
quaint basin
Tbf im curious why boxManager.addFriend needs to be on the main thread
bc i do player#getName
young knoll
```java
@EventHandler
public void onAsyncChat(AsncChatEvent event) {
...
(You can’t do event.setCancelled in that runTask :p)
thorn isle
that's thread safe i'm fairly certain but that's besides the point
thorn isle
(You can’t do event.setCancelled in that runTask :p)
well the whole thing was predicated on him not needing to modify the event
if you do need to modify the event, yes, you have to use the sync event
quaint basin
that's thread safe i'm fairly certain but that's besides the point
is not
public GameProfile gameProfile;
sly topaz
what
quaint basin
public GameProfile gameProfile;
It's not final, so there's no guarantee of visibility between threads
or volatile or smth
sly topaz
game profile isn't changed in a session unless a plugin explicitly does that, and they won't be able to do that on the same tick a chat message is sent
thorn isle
.
sly topaz
so, you're not really seeing the picture there
sly topaz
why
quaint basin
why not
thorn isle
what is it going to read if not the correct data?
a player's game profile, especially the name, is not changing mid-session; if it is, you have bigger problems to worry about than a marginally-stale read
sly topaz
because it complicates a system for no reason, you're trying to introduce thread safety concepts on a platform that is inherently single-threaded
there are cases where you do have to worry about it, this isn't one of them
quaint basin
a player's game profile, especially the name, is not changing mid-session; if it...
It shouldn't change mid-session, but it's not declared as final, so the JVM model handles the variable differently
sly topaz
It shouldn't change mid-session, but it's not declared as final, so the JVM mode...
it does not
a read is a read, it doesn't matter whether it is final or not in that sense
only thing final assures is that the value won't change after construction
thorn isle
you are oversimplifying the JMM
a field being plain and nonfinal does not mean 1:1 that it is unsafe to read
sly topaz
I think I've said it before but if you're looking to learn about concurrency, a plugin isn't the best project for it
thorn isle
for example, an object that is itself "published", i.e. made available via volatile memory semantics, and whose fields are written to once before that (i.e. are effectively final) is guaranteed to have visibility for those fields on all threads
unless you twist yourself into a pretzel to try and get around it, this will apply to this gameprofile field
young knoll
I do like pretzels…
thorn isle
the only special treatment final gets is a memory barrier at the end of the constructor
quaint basin
only thing final assures is that the value won't change after construction
what i mean is: variable X = 50. I change the value of X to 100 in thread X. from here, 1 year later, when reading in thread Y, the value of X can be either 50 or 100
If the game profile is changed mid-session, the thread may never see the change
thorn isle
well, it isn't going to be
sly topaz
that is not how it works
quaint basin
And I gave the example of the player's name, but I have vars of my own objects as well
sly topaz
what i mean is: variable X = 50. I change the value of X to 100 in thread X. fro...
a data race can't happen years later, or heck even miliseconds later is often already too late for that to happen
quaint basin
a data race can't happen years later, or heck even miliseconds later is often al...
Do you have a 100% guarantee of that?
I know it's not common, but theoretically it could happen
quaint basin
wdym thread local
thorn isle
i recommend this for reading if you're interested in how the JMM actually works
at any rate, i'd probably make cache a concurrent or synchronized collection and wrapping the rest in a sync task; it will still perform considerably better than scheduling the entire event to main
sly topaz
the issue here is that you're too fixated on the idea that happens-before relationships are only defined by the JMM
thorn isle
regarding the profile changing mid session, if something does that, you have bigger problems, e.g. the uuid could change under you and your cache no longer works
it is not, i don't think, something you should waste energy worrying about
quaint basin
regarding the profile changing mid session, if something does that, you have big...
But I have a object like this too:
Object {
private String var
setVar(String var) {
this.var = var
}
}
And I change the value of var in the main thread at any time.
quaint basin
My point is whether synchronizing the entire event wouldn't be more resource-intensive than using structures like concurrenthashmap, etc
thorn isle
it would be
slender elbow
oh man
sly topaz
other option that would solve this issue is probably to just not use the chat for this lol
slender elbow
I feel like Phil in Groundhog Day
sly topaz
feels like a command would be simpler
thorn isle
is this some bukkit conversation api replacement
quaint basin
other option that would solve this issue is probably to just not use the chat fo...
The best option would be dialogues
quaint basin
is this some bukkit conversation api replacement
yea
I did this a while ago
Is Redis more suitable than MySQL for synchronizing items between servers?
dawn flower
Does spigot have a way to register enchants to the enchantment registry like papermc does?
mortal vortex
with nms u get sum like:
MappedRegistry<Enchantment> registry = (MappedRegistry<Enchantment>) BuiltInRegistries.ENCHANTMENT;
unfreeze it, build and register ur data, then refreeze.
dawn flower
unfreeze it, build and register ur data, then refreeze.
alright, thanks
mortal vortex
allahu akbar
buoyant viper
https://pastebin.com/hnvm4bg7 is this enough datapoints for tracking identity?
buoyant viper
wait... i forgot network interfaces
mortal vortex
incredibly based
mortal vortex
wait... i forgot network interfaces
lets make Spigot's first invasive telemetry API and sell it to the NSA for 10 gazillion sheckles
🇮🇱
buoyant viper
lets make Spigot's first invasive telemetry API and sell it to the NSA for 10 ga...
TRUE!
echo river
I'm gonna explode my computer
getFacing and setFacing are not the same
LIKE
setFacingDirection(BlockFace.NORTH) sets the direction to the south
😭
(1.8.8 is peak)
obsidian sapphire
I do Minecraft server setup & fixes (Paper/Spigot). Plugins, permissions, errors. Cheap starting prices. DM me if you need help 👍
wet breach
setFacingDirection(BlockFace.NORTH) sets the direction to the south
right it is done as if you are looking at the block
so if its facing north then you are looking at it from the south
echo river
yeah but its so misleading
also its the opposite of what f3 says
like wtf
also u cant setData for a blockFace with a directionaldata
because why not
if (this.getDirection() != null && materialData instanceof Directional) {
Directional directionalData = (Directional) materialData;
RelativeBlockDirection relativeDirection = this.getDirection();
AbsoluteBlockDirection absoluteDirection = direction.getAbsoluteWith(relativeDirection);
directionalData.setFacingDirection(absoluteDirection.toBlockFace());
}
// Then updates the block
blockState.setData(materialData);
blockState.update();
}
and I have to use setRawData but its fine ig
// If the block is directional, adjusts its facing
if (this.getDirection() != null && materialData instanceof Directional) {
Directional directionalData = (Directional) materialData;
RelativeBlockDirection relativeDirection = this.getDirection();
AbsoluteBlockDirection absoluteDirection = direction.getAbsoluteWith(relativeDirection);
directionalData.setFacingDirection(absoluteDirection.toBlockFace());
}
// Then updates the block
blockState.setRawData(materialData.getData());
blockState.update();
}
you can submit a pr for it, but keep in mind setfacing and getfacing have existed long before all this new stuff now in the game lol
so during that time it made sense
for those of us who been around the api for a long time, still makes sense XD
fierce cipher
this is my challenge i doubt u can do it btw, its a yes or no and u have to do it in a week 🙂
Can you make a server that runs Minecraft 26.2 Snapshot 4 by remaking paper to support 26.2, and supports paper & purpur 1.8.8 - 1.21.11 plugins by automatically remapping the plugins? It should support eaglecraft and bedrock edition (Port 19132), and should work with bungeeguard.
its ok if u cant btw
wet breach
?services
undone axleBOT
If you wish to request or offer development/art/building/administration services, please do so at https://www.spigotmc.org/forums/services-recruitment-v2.54/
fierce cipher
does ANYONE i mean ANYONE know how to do that
lol ik u guys prob cant
wet breach
I could, but you are not paying me to do so nor could afford it
fierce cipher
WAIT
u can
i dont actually want u to do it i just want to know if u can do it @wet breach
can i dm u??
wet breach
i dont actually want u to do it i just want to know if u can do it <@22780086302...
so first, 26.1.2 paper has as an experimental build so there is no need to remake anything
echo river
Crazy tech
wet breach
so the answer is yes its doable
fierce cipher
so first, 26.1.2 paper has as an experimental build so there is no need to remak...
look at my dm?
wet breach
I don't read DM's
fierce cipher
can u do it in one week LOL
fierce cipher
@wet breach can u do it in a week
echo river
This is embarassing
slender elbow
lol
wet breach
I mean, what is there left to do when there is builds for that version you want?
echo river
This guy...
echo river
I mean, what is there left to do when there is builds for that version you want?
He doesnt understand anything prob
fierce cipher
so u can do it?
wet breach
so u can do it?
I guess? there is nothing left to do
echo river
Lmaoo
fierce cipher
its the newest snapshot r u ok?
echo river
LOL
wet breach
and there is paper versions for it
echo river
He pisses me off so much
wet breach
what about it?
echo river
The work is done @fierce cipher
fierce cipher
OH BET do u want dev in a diff server?
echo river
What ?
wet breach
o.O
fierce cipher
you will most likely get it
echo river
Hes sped omgmfgm
wet breach
they are sounding like some kind of bot lmao
fierce cipher
me???
echo river
Chatgpt with mishrioms
slender elbow
u have to do it in a week
i dont actually want u to do it
fierce cipher
>u have to do it in a week
>i dont actually want u to do it
yo ?
fierce cipher
YO WHAT
wet breach
doubt those who actually could be offended are not here
but I am not sure what the staff would think
echo river
Hahahahaha
fierce cipher
forstalf do u want dev??
echo river
Lolol
wet breach
you can't afford my rates
echo river
Ask for 50k a month
echo river
you can't afford my rates
What are they (jst asking)
Yo frostalf can u work 12h a day for free ?
For me plz
And answer my dm quick