🤖│community_dev

[floral crane]

https://github.com/WootingKb/wooting-rgb-sdk/blob/master/src/wooting-usb.c#L416C18-L416C18

[quick bough]

Well, judging by the code that uses keyboard_handle_array, I can forsee some issues with multiple keyboards being used.

So, the cleanup code seems to work "fine," as in it just cleans out the slot that the keyboard was using, so it doesn't reallocate/reindex other keyboards.

However, when a new keyboard is inited, it looks like it takes connected_keyboards as an index

Although it doesn't seem like this variable is ever decremented, so... probably an array out of bounds if you disconnect an reconnect a device 10 times.

But it should at least mean the device ids are steady.

[floral crane]

I asssume the max is at least 255, if not way higher

[floral crane]

hmm that might need to be increased or decremented lol

my keyboard disconnects constantly cause of my port

[quick bough]

Well, I think 10 is more than enough. The issue is that they don't "check which slot is free", instead they just keep incrementing until you have an array out of bounds

Ah wait, no, it looks like if any device is disconnected, it drops all devices.

So, when your disconnect callback is called, assume all device ids become invalid.

[floral crane]

ah great

I gotta constantly look for new device ids and link them to their serial numbers

hmm I can't do that though, to get sn I have to switch to that device first, it's not part of the usb meta

[quick bough]

Idk what you're doing, but it might be worthwhile to interface with the HID directly.

[floral crane]

that's basically what the rgb sdk does right

[quick bough]

Then you already have the serial number as its not abstracted away, you only have to send the reports yourself.

[floral crane]

I don't understand raw hid stuff, using the send command in rgb sdk is already more than I understand

[quick bough]

Well, there you are already sending a report yourself?

[floral crane]

yeah, that's through the rgb sdk though

[quick bough]

I forgor I have my Razer Huntsman V2 Analog connected right now instead of a Wooting® Analog Keyboard™ so I can't make a quick sample, but the HID report format is basically the same as that API call expect the first parameter goes last.

[floral crane]

I'm working out how to read the serial number with command 3 right now https://gist.github.com/BigBrainAFK/0ba454a1efb43f7cb6301cda8838f432

[quick bough]

I think this is a good example

[floral crane]

well that's good, if I ever want to switch to raw hid it shouldn't be too hard

[quick bough]

but yeah I guess the GetSerial command will also work if you can't get it from the HID itself

I think I'm gonna plug in my Wooting again, this Razer keyboard sucks, even if it is full analog 😠

[floral crane]

I wonder what this is

[quick bough]

What are you looking at?

[floral crane]

that's the serial number buffer

but how does it convert to "A02B2106W031H00418"

[quick bough]

Noo you just leaked your serial number

[floral crane]

oh no not my serial number

someones gonna seal my keyboard

[quick bough]

It's like your keyboard's SSN smh

[floral crane]

but I gotta figure out how to convert so people can compare to Wootility and figure out what keyboard it is if they have multiple of the same keyboard for some reason

[quick bough]

I'm just looking at it now myself, and it is perplexing

[floral crane]

decipher

function formatSerialNumber(i) {
        const o = new SerialNumber(i)
          , {SupplierNumber: s, Year: st, WeekNumber: dt, ProductNumber: ft, RevisionNumber: mt, ProductId: xt, Stage: bt, UniquePCBDesign: Ct, MinorRevision: St, Variant: $t} = i
          , _t = st.toString().padStart(2, "0")
          , kt = s.toString().padStart(2, "0")
          , jt = dt.toString().padStart(2, "0")
          , Nt = ft.toString().padStart(2, "0")
          , Mt = xt.toString().padStart(5, "0")
          , Ht = ProductionStageToString(bt)
          , Wt = o.hasOwnProperty("Variant")
          , Gt = o.hasOwnProperty("UniquePCBDesign")
          , Vt = o.hasOwnProperty("MinorRevision")
          , zt = Gt ? `T${Ct.toString().padStart(2, "0")}` : ""
          , Jt = Vt ? `${St.toString().padStart(2, "0")}` : ""
          , Yt = Wt ? `S${$t.toString().padStart(2, "0")}` : "";
        return ["A", kt, "B", _t, jt, "W", Nt, zt, mt, Jt, Yt, Ht, Mt].join("")
    }

[quick bough]

That makes some amount of sense

[floral crane]

what's SerialNumber though

[quick bough]

Well, I found bufferToSerial, which provided me with enough info to parse out the first chunk

This seems to match the SN fully, but it's with a bit of guesswork, so may not be perfect

r.skip(3); // D0 DA 03
r.skip(2); // no clue
uint16_t supplier_num;
r.u16(supplier_num);
uint8_t year;
r.u8(year);
uint8_t week;
r.u8(week);
uint16_t product_number;
r.u16(product_number);
uint16_t revision_number;
r.u16(revision_number);
uint16_t product_id;
r.u16(product_id);
uint8_t production_stage;
r.u8(production_stage);
std::cout << std::format("A{:0>2}B{:0>2}{:0>2}W{:0>2}{}H{:0>5}", supplier_num, year, week, product_number, production_stage ? 0 : 1, product_id) << "\n";

and I assume you're looking at the preview wootility which is where UniquePCBDesign comes in for the UwUs

So, that will likely impact the SN in some way

[floral crane]

use wooting_rgb_sys as rgb;

// https://gist.github.com/BigBrainAFK/0ba454a1efb43f7cb6301cda8838f432
const GET_SERIAL: u8 = 3;

fn main() {
    unsafe {
        let len = u8::MAX as usize + 1;
        let mut buf = vec![0u8; len];
        let response = rgb::wooting_usb_send_feature_with_response(
            buf.as_mut_ptr(),
            len,
            GET_SERIAL,
            0,
            0,
            0,
            0,
        );

        if response == len as i32 {
            println!(
                "A{:02X}B{:02}{:02}W{:02X}{}H{:05}",
                u16::from_le_bytes(buf[5..7].try_into().unwrap()),
                buf[7],
                buf[8],
                u16::from_le_bytes(buf[9..11].try_into().unwrap()),
                if buf[15] == 1 { 0 } else { 1 },
                u16::from_le_bytes(buf[13..15].try_into().unwrap()),
            );
        } else {
            println!("Error retrieving serial number");
        }
    }
}

[floral crane]

I'm trying to determine how I should interface device selection for rules, since the data type I'm storing is HashMap<String (serial number), u8 (profile index)>

Should I keep a list of all SNs ever connected even if not connected right now so you can always make rules for disconnected devices?

That would allow me to pre-populate a list of devices and profile indexes automatically, but if I don't do that I would have to have an add button and a popup to select which device to add

[quick bough]

That's more of a question regarding how you should design your app, which I think you would be best qualified to answer.

[floral crane]

This is going to be hard to do without a second device to test with, but if I understand correctly I need to figure out how to reset keyboard_handle and connected_keyboards, run wooting_usb_find_keyboard, then figure out a way to access connected_keyboards, iterate over that and send the command to get each devices serial number.

wooting_usb_disconnect resets but I can't access the number so I'll just have to loop until it returns false

Oh that's unfortunate, every index 0-10 is valid and returns the first device if one isn't found Select won't fail it will just silently fallback to the first device

[quick bough]

I have 2 wooting keyboards if you need me to test anything

[floral crane]

You can test this once the artifacts are uploaded by actions (in around 30-40m) https://github.com/ShayBox/Wooting-Profile-Switcher/releases/tag/2.2.0 This still doesn't support Wootility Beta and the new RGB effects system, but it should support multiple devices and the UwU

[strong siren]

i have 3 boards and use them all at the same time with artemis for testing frequently

when that feature was first added it was buggy but it seems good now

[floral crane]

The only problem I see is that every rule updates every device, but I can add a toggle to skip devices when I update to support Wootility Beta / RGB effects whenever that becomes stable

[quick bough]

30 to 40 minutes?!?!

I have C++ projects with thousands of source files and we batch build them in 1/10th of that time

[floral crane]

I have release builds set to single thread compilation to reduce file size as much as possible, it could be done faster,

[quick bough]

What does multi-threading have to do with file size?

[floral crane]

but this project is the largest dependency tree out of all my projects and takes the longest, I might have to remove that if it takes much longer

codegen-units = 1 # Compile crates one after another so the compiler can optimize better

https://github.com/johnthagen/min-sized-rust?tab=readme-ov-file#reduce-parallel-code-generation-units-to-increase-optimization

[quick bough]

Sounds like a weird Rust issue

I would die if I had to compile my projects with like 2000 source files on only one thread

some of these source files, no joke, take 20 seconds

but my i9-13900K gets the gamut done in under 1 minute

[floral crane]

Those kinds of optimizations exist with gcc too, but it doesn't really make a big difference I just enable it on all my projects cause compile time on actions doesn't matter

it doesn't apply to debug builds so it compiles faster on my computer

I could figure out how to setup caches to speed it up

[quick bough]

So, I plugged in the second wooting, seems I needed to restart the program for it to recognise it

[floral crane]

yeah it only checks on startup

It would take more work than it's worth to re-initialize the tray

it only took 2m to build on my computer

[quick bough]

That is without optimisations?

[floral crane]

without any file size optimizations

[quick bough]

Who cares about file size, CPU time is important

[floral crane]

compilation time doesn't really matter

30 longest to compile dependencies

[quick bough]

So, I've set up a thing to make it switch profile when GTA5.exe is focused, and it does seem to work, but when I unfocus GTA V, the second keyboard doesn't seem to have its active profile reset.

[floral crane]

The program will use the active profile at the first start (when it creates the config file) as the fallback profiles for the keyboard when no matching rule is found, I don't have a ui to edit it but it's in the config file and uses the same format as rules. you can also create a rule with wildcard in all the fields to achieve the same effect, I believe it has to be the last rule (bottom) but it might be top

[quick bough]

It looks to me like the "WootDevResetAll" command was not sent

I did it manually and now the keyboard is back to normal

[floral crane]

It should have sent it

[quick bough]

Well, it might be that the keyboard never gets the command to swap back the profile either

Idk what to say, but it's definitely a consistent issue

Only one of the keyboards is reset after defocusing GTA V

[floral crane]

oh hmm, what's your fallback config setting

[quick bough]

My what, where

[floral crane]

it might have populated the first time when you didn't have both plugged in and doesn't have the 2nd

Edit > Open Config File

[quick bough]

Ah, this one didn't get it populated.

[floral crane]

Ah yeah that's a flaw, I need to add a UI to edit the fallback and a toggle to skip devices

you can empty the fallback_device_indices and open the program with the profiles active that you want to be the fallback to fix that for now

[quick bough]

I have manually added the second device to the fallback_device_indecies array with a value of 0

now the program crashed on defocus and neither of the keyboards were reset

[floral crane]

I also just realized I forgot to update the version number on the release files lol

[quick bough]

can't reproduce it now, seems to be fine again

weird stuff

wasn't Rust supposed to not have random crashes

[floral crane]

I'm sure there was a reason, there's just no logs to find out why unless it's run from a terminal. probably forgot the comma and couldn't parse it and reset the config

[quick bough]

I know JSON isn't a very human-friendly format but I hardly pass as one anyway 😛

That is to say, I definitely did not mess up my commas.

Welp, whatever, it is working perfectly fine now

A weird thing is that my Wooting Two seems to be flashing at 100% brightness shortly before its profile is switched

but I doubt that's a you-problem

[floral crane]

I'll add that to my list of things to add with the next update when wootility updates

that's likely the send sleep delay, I use a higher safe value than wootility probably does, you can lower it or make it 0, I've never had an issue at 0

[quick bough]

Set it to 0, but still the same thing, just faster I guess

but, well, this wooting two is beaten to death anyway

ah wait I just realised, the second profile has a different brightness setting than the first one

so yeah not your issue

[floral crane]

I removed the single codegen unit optimization from the re-release, I wonder how much faster it will build

[quick bough]

weird tho how it retained the wrong brightness then

[floral crane]

brightness can be changed per-key, pey-profile, and per-device, and wootility previews it in-memory without writing it you press save. but I never write to the device in WPS

[quick bough]

yeah maybe some weirdness with that previewing

Seems your app somewhat-gracefully handled one of the keyboards being unplugged as well

Only issue is no tray update but yeah

[floral crane]

the devices will never disappear from the tray or elsewhere, I store the serial numbers in a list so you can configure unplugged devices

[quick bough]

ah, hmm

[floral crane]

clicking the button in the tray just does nothing but print a not found message in the console

[quick bough]

what are they supposed to do anyway

they are checked on, but I click on them and they stay on

[floral crane]

you can click the device if it's connected to select it and then the profile selection below applies to the selected device

hmm

[quick bough]

Some minor usability tweaks I would do is maybe sort the keyboards in descending order, since higher = newer = more likely to be what the user wants to use

Also maybe display the actual name of the keyboard instead of the SN

it's highly unlikely the user will own 2 of the same model, but more likely they will own different models

[floral crane]

yeah I want to show the model name in a shortened form and have a toggle to show the SN instead. it also doesn't pull the profile names from the other devices since wootility only added multiple devices in beta

I might change the tray to instead have a label showing the device and a subset of profiles under it, instead of two selections, double the list and show both sets of profiles

[quick bough]

soon I will own another wooting + a wooting UwU

then I will plug them all in and watch the USB ports in my mobo die a miserable death

that is to say, I think users using multiple keyboards in general is highly unlikely

maybe more likely-ish with the UwU

if you can call that a keyboard

more of a macro pad, or "Osu! controller"

[floral crane]

the uwu is the only reason, even wooting didn't support it until now in wootility

[quick bough]

Although I can't help but think of Tom Scott's Emoji Keyboard

Imagine if he built this nowadays with the Wooting Analog SDK

it would be a breeze, you have the SN for reference

get proper reports for each key pressed

then just use SendInput from WinAPI, and you're on your way

[floral crane]

Looks like build time was cut in half, back down to before I switched to tauri v2 or added game scanning

[quick bough]

Now rewrite it in C++ and weep

[floral crane]

it was

https://github.com/ShayBox/Wooting-Profile-Switcher/tree/original

c anyway

[quick bough]

[floral crane]

what's that for

[quick bough]

A big project 😛

What's even sadder is that this is Windows build times

If I could switch to Linux, it would probably only take 20 seconds

[floral crane]

10x slower on actions, it takes most of the build time to just install tauri-cli on actions

[quick bough]

Stuff compiles so incredibly fast on Linux

[floral crane]

linux is always the first build to finish

windows after, even though I build twice (portable is a debug build) and macos last

lots of shared libraries on linux I guess

[quick bough]

the shared libraries are such a bane to me

instead of relying on a dynamic linker, I basically just dynamically link myself on linux

because various complications

.<

[floral crane]

I really wish Wootility wasn't a single large minified javascript file. I see it's using a deviceId variable as the key for storing data connected to a keyboard, but what is this? I can't figure out how it's created. If I had to guess it's either the HID ID (I hope not) or an encoded form of the serial number (this seems to be the only reliably unique identifier)

[quick bough]

Let's just bribe Simon to have him hand us the Wootility source code

I'm actually a bit surprised the Wootility isn't open-source since it's just a "glorious" bundle of JavaScript...

So, severely more trivial to reverse-engineer than, say, a DLL/SO written in Rust...

[floral crane]

I give up, I found the legacy serial number code but not the device id code

[quick bough]

I don't see why device id would not be the same as the serial number

[floral crane]

I don't either, maybe it is somehow but I can't find it. mine is 241831621

[quick bough]

Ah hmm interesting

[floral crane]

it seems to be the same if I wipe the data, I'm gonna try changing ports between

still the same

[quick bough]

What's weird to me is that the device id is not a string, because all the code I can find relating to it calculates a string

[floral crane]

same if I restore firmware to pre-beta and open the beta wootility

it is a string, in json

[quick bough]

I mean, it is clearly a number?

[floral crane]

deviceId is the variable name used in the local storage, for some reason it's stored as a string but it's really a number

[quick bough]

Oh, wait

lmfao

Look at 241831621. Here, 21 is the year it was made.

        class Mc extends Bn {
            constructor(e, t) {
                super(e, t)
            }
            async getlayoutType(e=!1) {
                if (e || void 0 === this._layoutType) {
                    const e = await this.getDeviceConfig();
                    this._layoutType = e.getRawKeyboardLayout()
                }
                return this._layoutType
            }
            async getDeviceID() {
                if (void 0 === this.deviceID) {
                    const e = await this.getSerial();
                    this.deviceID = "" + this.keyboardType + e.ProductId + e.ProductNumber + e.RevisionNumber + e.WeekNumber + e.Year
                }
                return this.deviceID
            }
        }

And 16 is the week. Very similar to serial number...

[floral crane]

Ah so it is the serial

why did I not try to search for getDeviceId

[quick bough]

Except it also includes the "keyboard type"

(Whatever that means)

[floral crane]

probably full, 80, uwu etc

[quick bough]

My device id seems to be 32504413821 for a Wooting Two HE

I definitely recognise some parts of it from the serial

[floral crane]

WOOTING_USB_META shows the type, mine is 2

Type: 2 | 418316 | Week: 21

418 is at the end of my serial's 5 digit number 00418 (padded)

[quick bough]

My Wooting Two HE is definitely not a DEVICE_KEYBOARD_60

[floral crane]

Type: 2 | H: 418 | W: 31 B: XXX6 | Week: 21

[quick bough]

21 is year, not week

[floral crane]

wait yeah why is yours 3 when that's the 60HE

I guess 3 is all HE not just 60

[quick bough]

Well... no...

I think it must be something else

[floral crane]

hmm

[quick bough]

I think it might be from this:

[P.WOOTING_ONE = 0] = "WOOTING_ONE",
P[P.WOOTING_TWO = 1] = "WOOTING_TWO",
P[P.WOOTING_LEKKER = 2] = "WOOTING_LEKKER",
P[P.WOOTING_TWO_HE = 3] = "WOOTING_TWO_HE",
P[P.WOOTING_60_HE = 4] = "WOOTING_60_HE",
P[P.WOOTING_60_HE_ARM = 5] = "WOOTING_60_HE_ARM",
P[P.WOOTING_TWO_HE_ARM = 6] = "WOOTING_TWO_HE_ARM",

Is your keyboard a Wooting Lekker?

[floral crane]

Ah just a different incremented list I guess

yeah
Type: 2 | PID+PIN: 418 | REV: 3 | Week: 16 | Year: 21

[quick bough]

Then that makes sense

[floral crane]

well this is good, idk why the sn wasn't used directly though lol

[quick bough]

Not unique enough maybe

[floral crane]

I think it's less unique than before, it's missing one of the numbers from the SN

no nvm it's every number, just in a different order and without padding

[quick bough]

The lack of padding is what's questionable to me, because in theory there could be ambiguities due to that

like, as we saw, my device id is a bit longer than yours

but, meh, I guess it's "good enough, ship it."

[floral crane]

🤔 A02B2106W031H00418 BXX06 is not 16 for the week

[quick bough]

No, it's 6 for the week. Lack of padding, as I said.

This format is unfortunately unparsable due to that as well, unless you do some sort of "least vexing parse" magic.

[floral crane]

I'm just gonna write a serializer to convert

[quick bough]

You can convert serial number to a data structure and that data structure to the device id

but device id to a data structure

[floral crane]

oh the keyboard type probably means board cause of the B after it

W for week H for something related to 031

[quick bough]

Huh? The thing between A and B in the SN is supplier number.

[floral crane]

but that's the keyboard_type

[quick bough]

Please don't confuse the SN for the device id

[floral crane]

oh I see

device id is missing the supplier number and using its own keyboard_type which happens to be the same number for me

#[derive(Clone, Debug)]
pub struct Device {
    model_name: String,
    supplier:   u16,
    year:       u8,
    week:       u8,
    product:    u16,
    revision:   u16,
    product_id: u16,
    production: bool,
}

#[derive(Clone, Debug, Display, Eq, Hash, PartialEq)]
pub struct DeviceID(String);

impl From<&Device> for DeviceID {
    fn from(device: &Device) -> Self {
        let keyboard_type = match device.model_name.as_str() {
            "Wooting One" => 0,
            "Wooting Two" => 1,
            "Wooting Two Lekker Edition" => 2,
            "Wooting Two HE" => 3,
            "Wooting 60HE" => 4,
            "Wooting 60HE (ARM)" => 5,
            "Wooting Two HE (ARM)" => 6,
            &_ => 7,
        };

        Self(format!(
            "{}{}{}{}{}{}",
            keyboard_type,
            device.product_id,
            device.product,
            device.revision,
            device.week,
            device.year
        ))
    }
}

#[derive(Clone, Debug, Display, Eq, Hash, PartialEq)]
pub struct DeviceSerial(String);

impl From<&Device> for DeviceSerial {
    fn from(device: &Device) -> Self {
        Self(format!(
            "A{:02X}B{:02}{:02}W{:02X}{}H{:05}",
            device.supplier,
            device.year,
            device.week,
            device.product,
            device.production as u8,
            device.product_id
        ))
    }
}

[floral crane]

So serial number doesn't use the revision number or keyboard type, and device id doesn't use the supplier number or production stage

🤔 which one is more unique

[quick bough]

One you make yourself 😛

But I'd def. prefer SN because it's parsable

And also something the user can easily look up themselves

[floral crane]

yeah I'm gonna use SN with FromStr to DeviceSerial, and just have a From/Into DeviceID from Device so I can access the data in Wootility, but because it lacks all information I can't directly create a Device from the serial number, I have to find it by scanning devices

[quiet root]

its the exact order of releases

[quick bough]

If I had to guess, UwU is 7 and UwU RGB is 8 😛

[floral crane]

Is there an ETA for Wootility Beta to be stable? and do we know the exact strings and ids for the UwU?

The RGB SDK feature/uwu branches has the strings, but does anyone have an UwU? can you open Wootility Beta, press Ctrl + Shift + I, and go to the Application > Storage > Local Storage > file:// and tell the first number of your device id

[lime charm]

Oh hell yeah there's a .NET wrapper

Now I can easily play around with the keyboard as a C# developer

[signal gust]

if you're still wanting this, 7

been a few hours so i dunno but i thought i may as well

[floral crane]

is that the uwu or uwu rgb

[signal gust]

uwu

non rgb

though my id seems to be a bit longer than what you had highlighted in the image

14 digits if i can count correctly

dunno if thats something you already knew

[floral crane]

the device id length is dynamic unlike the serial number which pads leading zeros to a fixed length

[signal gust]

ah okay

[lime charm]

I really have to learn Rust one day

[floral crane]

If I had to guess your product id is longer than mine (3), and likely your week is 2 digits intead of 1

Rust is great 🦀

[signal gust]

rust truly is great

[lime charm]

currently learning C and ASM at uni

[floral crane]

oh no

[signal gust]

also learning C at uni and probably asm soon

also learning python at the same time but i already know that and the lectures are just free sleeping time

[lime charm]

Ergh I don't like Python. I never use it anyway

[signal gust]

for me python is good for scripting but i wouldnt use it for anything "big"

[lime charm]

Using it for anything big is misusing it lol

[signal gust]

not a huge fan but using it more is making me appreciate it a little more

it does have some nice things here n there

[floral crane]

I started with Java but really learned JS, then TS, PY, GO, and RS but not in school. I've tried C/C++ but it makes me hate programming

[lime charm]

JS/TS makes me cry

[signal gust]

i love js

unironically

liek yeah its stupid

but its just nice sometimes

though i definitely would always go to TS now

[floral crane]

I can deal with TS, but JS is just awful unless it's a small amount of code

[lime charm]

I hate that there are no object types in runtime

[signal gust]

yeah going to js feels so bad now

[lime charm]

but even ts has no types at runtime

no reflection

[signal gust]

i dont really find it to be an issue honestly

[floral crane]

TS is just a tool to make JS less bad

[signal gust]

to make it actually usable in big projects

dunno how anyone survived without it

[floral crane]

if you're able to use something like deno that runs ts why not just use something else entirely. leave js to die in the browser

[signal gust]

bun!

or deno yeah

but bun is fun

[floral crane]

yeah bun, never tried it

[signal gust]

its uh

its pretty nice i guess

rn its nothing super special (in my opinion/for what i'm doing anyway) like people are making it out to be

but its faster at some things and its mostly a drop in replacement for node

[floral crane]

I need to learn how to write wasm with rust for websites

[lime charm]

I want Blazor to get really good :)

[floral crane]

I never really learned C# or F#, I tried it when dotnet core came out and it's actually a usable language for more than just Windows programs. but I still don't have a reason to learn and use it over Rust, even for the windows api

[strong siren]

yeah blazor is still kinda poop imo

[lime charm]

I have been using C# for yeaars now but recently toned it down a bit

Because of Uni

[signal gust]

i think C# is pretty nice now

or has been for a while

[floral crane]

It's nice, certainly more enjoyable than Java

[signal gust]

i recall hearing it wasnt as nice

[lime charm]

Yeah it's fast and cross-platform now

[strong siren]

in the framework days no, not really

and open source!

[lime charm]

mostly the binary sizes in the browser

[signal gust]

i also feel like java isnt that bad compared to how much hate it gets but its certainly not amazing

[lime charm]

no reason to use Java when C# exists though :)

[floral crane]

Java isn't that bad but it's also only use for like Minecraft these days, so why learn it over anything else unless you need to

[signal gust]

how do i write my minecraft mods in C# 😭

even then just use kotlin

smiley face

[strong siren]

i'm currently trying to do a thing with opencv + wasm + blazor and had to use .NET 8 preview for some things, it feels like it has lots of rough edges still

[lime charm]

Just uhh. Imagine it working

[floral crane]

Rust to JVM when

[lime charm]

Sadly yeah

[floral crane]

Minecraft mods written in Rust 🤔

[lime charm]

Minecraft written in Rust!!!

[signal gust]

yes

[lime charm]

What about Zig? Bun uses that

[signal gust]

zig is apparently cool but i havent looked into it

[lime charm]

Apparently it deals with memory in a different way than Rust

Its biggest downside is that you cannot say you're writing Rust while using it

[signal gust]

truth

i think zig memory management is manual but you can choose the allocator yourself and stuff

[lime charm]

I heard it's memory safe though

[signal gust]

ye

[lime charm]

But how it achieves it without checking array bounds etc I don't know

C# and Java spends CPU instructions checking bounds

[signal gust]

dunno much about it i dunno either

[orchid falcon]

zig is not realy memory safe like rust
sure it has some syntax to make is safers (slices and optionals being prety much the defauld)
but it also doesent have the restrictions on memory rust has
but what is nice about zig is the no hidden controll flow
errors have to be handeld you cant ignore them

[lime charm]

but not like checked exceptions in java?

[orchid falcon]

from quickly looking what checked exceptions is its pretty much the same
https://ziglang.org/documentation/0.11.0/#catch
var value = functionThatCanError(); -> compiler error
var value = try functionThatCanError(); -> keep throwing the error
var value = functionThatCanError() catch { //deal with error };

[floral crane]

https://github.com/ShayBox/Wooting-Profile-Switcher/releases/tag/2.3.0 I released a pre-release for use with Wootility Beta and it'll be marked as latest when that becomes stable

similar in rust, Result and Option types can use ? to pass to the current if its also the same Result or Option return or you can match/if check them for Ok/Err Some/None to handle the errors or let _ = to ignore the response

[lime charm]

they just move it to a try instead of throws at fucking declaration - which makes it clearer where it's being thrown. Neat

[quick bough]

This is in the beta wootility:

    o[o.WOOTING_UWU_NON_RGB = 7] = "WOOTING_UWU_NON_RGB",
    o[o.WOOTING_UWU_RGB = 8] = "WOOTING_UWU_RGB",
    o[o.WOOTING_60HE_Plus = 9] = "WOOTING_60HE_Plus",

I guess they kinda leaked the Wooting 60HE Plus, whatever that is supposed to be >.<

The base PID of the Wooting 60HE Plus will be 0x1320

[floral crane]

good to know, I'll add that to the wootility aur package and wps

[quick bough]

What I don't understand is why Wooting is deciding to focus more on 60% keyboards than TKL keyboards.

Then again, I never had to manage a company that deals with physical products.

[floral crane]

I don't like anything but full keyboards but I prefer the size with a delete button and arrows. In firmware key remapping and fn layer is the main reason I use this keyboard, everything else is bonus

[quick bough]

Well, I'd also never buy a keyboard without a numpad, but it's not like we 2 people represent 100% of the market. 😄

Wooting's software is also quite nice. Recently I did try the Razer Huntsman V2 Analog because it is a full-analogue keyboard, but I quickly realised how much customisation that Wooting was offering I now no longer had, and the Wooting Two is quite compact in comparison, so it's overall a nicer keyboard.

[floral crane]

Wooting has the best firmware and software even if it's not open source

[quick bough]

Well, Warframe did do funny RGB effects on the Razer keyboard... but that's just a lack of support. 😄

It's a shame there's no universal keyboard RGB thingy

[torpid fox]

When using the analog SDK, every so often my mouse becomes unusable for a second or two, anyone have a clue as to why this happens?

it's quite annoying, and also happens when using software others wrote using the SDK (like the input display)

[quick bough]

I noticed the Analog SDK has some weird lag issues. I ended up rewriting a part of it in C++, and no more lag: https://github.com/calamity-inc/universal-analog-plugin

[floral crane]

plug them into different usb controllers if you can

[torpid fox]

I'm well aware that you did that because you've mentioned it like 4 times to me now and it's not going to help me as I've also mentioned several times

[quick bough]

Well, it's my fix 😛

Have you tried it, does it work?

[torpid fox]

it doesn't answer my question though

I haven't tried it and it doesn't answer my question

[quick bough]

I have no idea why the lag is there, no.

[floral crane]

i dont use analog, I disable it so i can use my gamepad in games

[torpid fox]

I'm using it on the one straight on my motherboard

same with my mouse

[quick bough]

It is also not relevant for my use-case of reinventing every wheel in C++. 😛

[torpid fox]

oh, we're talking about a different type of controller

[torpid fox]

okay then stop bringing up unrelated shit

[quick bough]

It is very much related tho

[floral crane]

put then in different sides front/back of the pc or far ends of the back ports. different onboard usb controllers

[torpid fox]

(pardon my french)

[quick bough]

But yeah I understand I don't have a solution that satisfies you

[torpid fox]

oh so we did mean the same thing
I can certainly try

[floral crane]

my wireless mouse has to be in the front with nothing else plugged in or it loses signag strength

[torpid fox]

I'm not using anything wireless

fyi

nope didn't help

[floral crane]

still

[torpid fox]

I tried it, didn't help

[quick bough]

I guess the USB controller might be overloaded, but I don't see why the issue would only be apparent when using the Analog SDK, then.

[torpid fox]

It definitely sounds plausible

I'm going to check device manager real quick to see if I actually have multiple controllers or if this is just motherboard smoke and mirrors

[floral crane]

things might be polling and recieving lots of analog data, imagine a whole keyboard of floats in milliseconds

[quick bough]

Not that I understand much about hardware, but I do think the keyboard always reports how far a key is pressed, and all the Analog SDK does is listen.

[torpid fox]

stuff works in events

I think

[floral crane]

idk how it works, never messed with analog sdk

[quick bough]

Your keyboard sends an update with all pressed keys and how far they are pressed when there is a change

It can support up to 16 keys pressed at a time in the report

So it would never send the full 100-something keys

Also it doesn't use floats, it uses a u8 with a value from 0 to 255

[torpid fox]

fine I'll try your analog sdk thing but I'm reading the source code first

[quick bough]

There's a lot of source code to be read, if you want to look at the full framework I used for this 😄

[torpid fox]

[quick bough]

But this is all the relevant code, I'd say:
https://github.com/calamity-inc/universal-analog-plugin/blob/senpai/main.cpp
https://github.com/calamity-inc/Soup/blob/senpai/soup/AnalogueKeyboard.cpp
https://github.com/calamity-inc/Soup/blob/senpai/soup/hwHid.cpp

[quick bough]

Well, alternatively you could be using namespace std::chrono_literals; so you can write std::this_thread::sleep_for(100ms);

[torpid fox]

that's not what-

Why are you sleeping 0.1 second?

[quick bough]

When no keyboard is detected

[torpid fox]

ooh okay.

Is calamity-inc/Soup your life project or something

[quick bough]

haha, no, I've only been working on it for max. 2 years

but we need HTTP, TLS, JSON, and stuff like that for our bigger projects

So it's all in there

[torpid fox]

well yeah but it's also a hot mess of bloat

[quick bough]

Not really

Dead code elimination exists

[torpid fox]

Well yeah but if I'm going to compile this myself

[quick bough]

If that's your goal, you'll need the Sun build system: https://github.com/calamity-inc/Sun

You can use the PHP scripts to "bootstrap" it

[torpid fox]

okay yeah no

[quick bough]

I mean you can probably build "Universal Analog Plugin" with Visual Studio, but I only used Sublime Text + Sun for it because it's simpler

[torpid fox]

Right now the choice is between reading more code than I currently have to or trusting you and I don't have any good reason to do either

[quick bough]

Would a PDB help you?

[torpid fox]

No I'm just done

[quick bough]

Alright then xD

[torpid fox]

you've been trying to shove this universal analog whatever down my throat too much so I'm just kinda done

[floral crane]

for a plugin mesnt to be used and integrated, having a different build system is an obstacle. ideally support all major systems however that may be

[torpid fox]

it's also written in a completely different language

[quick bough]

I support the only build system I consider worth using 😄

[torpid fox]

that doesn't mean we should

that's your choice and that doesn't impact me

[quick bough]

I mean, it's a single .cpp file, you can probably just invoke clang yourself to build it tbh

[torpid fox]

except I'm not doing it because I said I'm done with this shit

in the end it's not going to help anyway because I NEED to support the official SDK for my use case

no way around it

[quick bough]

You don't have to. I already said I know you're not satisifed with the solutions I have to offer.

I just personally use this plugin because a) I used the Razer analogue keyboard, and b) I don't like the lag with the official plugin.

[torpid fox]

You've been trying to push this shit down my throat 4 times now and haven't exactly sounded trustworthy

[quick bough]

Sorry it came across that way

[torpid fox]

Anyway I'm here to get my question answered and not to linger on your rewrite stuff because it is simply not an option

[quick bough]

FWIW, the only lag I experienced with the Analog SDK is related to init & deinit

[torpid fox]

For me it's consistent throughout runtime

[quick bough]

I was gonna say it might be related to your specific use case, but you did say you also had it with other apps using it?

You were trying to write Godot bindings, right?

idk if you figured out why Godot GC'd your plugin yet, but the constant deinit & re-init due to that might be a cause of lag then

[torpid fox]

I fixed the GC issue

[quick bough]

Nice

[torpid fox]

or rather

I worked around it

there are now unit tests and stuff that are queued for fixing this issue

[torpid fox]

but you did say you also had it with other apps using it?
Yes. Also happened with the input display.

[torpid fox]

Correct.

[quick bough]

Could I get a link to this input display? See if it works fine for me.

[torpid fox]

but the constant deinit & re-init due to that might be a cause of lag then
There was no constant deinit/reinit

[quick bough]

(Real life calling me now, I will test it once I get back.)

[torpid fox]

https://github.com/DjCrqss/Woot-verlay

I think it was this one

yeah

since you're gone for a bit I'm gonna redownload it and check whether this is the one and if the issue still happens

Yeah it also happens with this one

It very well might be some kind of bus bandwidth limit, because I have an old motherboard

And I’m not sure which USB ports go into what bus on my PC right now

Notes:

• this does happen on init as well but that could as well be a bandwidth thing because it has to send a bunch of stuff on init anyway
• I do have to actually connect to the application for it to start doing this

[quick bough]

nah, the init and deinit lag is "normal"

also did you just run the Woot-verlay-2.1.exe without checking the source or building it yourself? 😛

Anyway, just tested the Woot-verlay, seems to work fine on my machine, no mouse lag

[torpid fox]

but that's also not relevant because runtime lag isn't

I wonder if I can figure out whether this is a bandwidth limit using wireshark or something

[quick bough]

I'm like pretty sure the keyboard does always send those reports, it's just a matter of if your machine is listening or not

but it's also possible that my assumption about this is wrong

then the additional bandwidth could be an issue in processing your USB mouse

have you tried using the touchpad on your laptop while using the Analog SDK?

at least it looks a bit like a laptop idk

[torpid fox]

I am using a desktop computer

[torpid fox]

my computer is not visible in the shot

[quick bough]

I see

[torpid fox]

if that was the case, then I would be having issues when the SDK isn't running (I'm not)

[quick bough]

Well, I would ask if it's only an issue with the Analog SDK but you don't trust me with my evil C++ code

Have you tried polling the HID reports on usage page 0xFF54 yourself?

[torpid fox]

I have no reason to trust you and you lost the trust I am willing to give by default by being pushy

You're not making it any better right now

[quick bough]

Maybe you could also set up a linux virtual machine, attach the Wooting keyboard to the VM so you can use the Analog SDK there, and see if it impacts your mouse

also, if I were to peddle malware, why would I target you specifically? If I wanted to spread malware, I'd go for a bigger audience than just a single person. I also don't know you, don't think there's anything special about you to target you specifically, but whatever.

[torpid fox]

Before doing all of this I'd like to see if I am hitting a bandwith limit because it seems the most likely given my motherboard is 10 years old

[quiet root]

would be weird given all our keyboards can only do usb1.1 speeds

12mbit/s

[torpid fox]

hmmmm

[quiet root]

the usb2 speeds would meed a high speed coms chip that we simply dont have. so usb fullspeed only

[quick bough]

well, my motherboard didn't like me connecting 2 wooting keyboards into the same controller

[quiet root]

thats because of usb endpoints most likely

[quick bough]

but I think that was due to the power draw of the keyboard

[torpid fox]

USB endpoints?

[quiet root]

basically usb devices the OS sees

our keyboard shows up as several devices

mainly to facilitate certain functionality like mapping mouse buttons and gamepad output

[torpid fox]

I was about to mention a controller and mouse yeah

[quiet root]

i am unsure if linux has a concept of usb endpoint limits like windows does

[quick bough]

"our keyboard"

are you support staff at wooting?

[quiet root]

no

developer

[torpid fox]

cough

[quick bough]

Well, I don't see a tony on the "meet our team" thing

[quiet root]

https://wooting.io/about

all our staff is listed here

[torpid fox]

If the issue I am having is related to USB endpoints I feel like it would be happening constantly instead of specifically when the SDK is running

[quiet root]

correct

also it would just stop accepting new endpoint devices if that was the issue

but was more for sainan

their issue of 2 wootings is either a power limit or usb endpoints

[quick bough]

I can connect multiple wootings, it was just an issue of not on the same row

[quiet root]

depends on the mobo config. if the 2 ports share 1 controller and other ports another then it can still be endpoints

the endpoint limit comes from the controllers first and OS second

[torpid fox]

yeah, I was trying to dig deeper myself to see if it could be related to my issue. do you have any ideas as to why my issue could be happening?

[quick bough]

Well, the issue is that the wooting itself just turned off

So I think it just didn't get enough power

[quiet root]

keyboard doesnt turn on if it doesnt get enumerated

[quick bough]

It turned on

but it turned off again very quickly

[quiet root]

hm

[quick bough]

possibly the motherboard realising that the power draw is too much and hence forcing it off

anyway, thing of the past

[quiet root]

could be. back to the main topic @torpid fox do you have just 1 mouse connected to the pc?

[torpid fox]

Yeah, I could connect more for testing though.

[quiet root]

id try at least

rule out something to start off

[torpid fox]

Or at least, one physical mouse. The keyboard obviously counts as one

[quiet root]

how did you even find that ancient article from the first launch of keyboards

[quick bough]

I mean this thing 😄

[quiet root]

oh

yeah thats uh

"slightly" outdated

[quick bough]

I should've realised Wooting has a few more employees

[torpid fox]

Since it's literally on your front page that seems pretty important

[quick bough]

but also I have no idea what Wooting employees would be doing all day? xD

[quiet root]

we are sort of slowly trying to figure out what to do there although tbf those are the sort of main faces i guess

most of the livestreams are from 2 of those folks

[torpid fox]

Yeah I noticed that

[quick bough]

also I just realised the z-indexing is wrong when the page first loads and only gets fixed when hovering on it

[quiet root]

ive been on like 1 and that was that but yeah we arent exactly sure how to solve it as we have a few too many people to do the same thing again

[quiet root]

same but we didnt choose to fix it until we know what to do in general with the space

[torpid fox]

Mouse connected. I assume you want me to try to see if the issue still happens when trying to move the new mouse I just connected?

[quick bough]

Also, may I ask what you are developing for Wooting, Tony?

[quiet root]

yeah basically

[torpid fox]

Huh wait what now it isn't happening at all even with the other mouse

maybe I'm just getting unlucky but this seems weird

[quiet root]

no theres just something going on

[torpid fox]

Oh okay no it just happened

I was unlucky

[quiet root]

hm

[quiet root]

anything and everything at this point

[torpid fox]

Happened with the first mouse I mean

[quiet root]

wait so mouse 1 does it but mouse 2 doesnt?

[torpid fox]

Not sure yet

I'm not sure about mouse 2 yet because it suddenly isn't 100% consistent

[quick bough]

Well, someone leaked the 60HE Plus 😛

[quiet root]

the what?

[quick bough]

¯_(ツ)_/¯

It's something I found in the Wootility Beta

[torpid fox]

God damn you windows clipboard

[quick bough]

just thought it was interesting

[torpid fox]

I tried pasting a message link and it just failed

#🤖│community_dev message

[torpid fox]

okay this may actually be relevant because when I unplugged mouse 2, mouse 1 immediately starts acting up

[quiet root]

wait you have mouse 1 still plugged in

[torpid fox]

Should I unplug mouse 1?

You reckon I should plug mouse 2 in the port where mouse 1 used to be?

[quiet root]

well the same port might be good but in general switch out mouse 1 completely

[torpid fox]

oki

[quiet root]

as to eliminate mouse 1 from the test

[torpid fox]

Issue now happens immediately to mouse 2

[quiet root]

so more mice = less frequent?

[torpid fox]

...maybe I think so?

[quick bough]

or just that port not being nice to mice

[torpid fox]

That's what it seems like

[torpid fox]

Well it works fine normally

hm but I plugged the mouse into a different port and it hasn't happened yet

[quick bough]

silly hardware

[torpid fox]

it's specifically on a different area on the computer though that may be a different bus

it's the one I plugged the keyboard into earlier

(issue still happened)

Oh nope it just happened

specifically when I clicked

Maybe my motherboard is straight up broken

there are several other oddities

if this just ends up being a motherboard problem I'm going to be slightly pissed at myself

[quick bough]

Tony, would you happen to know if the firmware always sends the analog reports on 0xFF54 or if the host specifically has to ask for it? I don't really know how HID works below the OS level unfortunately.

Maybe Wireshark could answer this for me

[torpid fox]

Wireshark with usbpcap seems like a good way to find that out yeah

[quick bough]

Unfortunately I haven't managed to get usbpcap working on my machine

Oh well

but assuming it's that the reports are only sent when the host registers interests in them, then it might be your motherboard being overloaded then

[torpid fox]

Really? that's odd, it's always worked flawlessly for me

[quick bough]

Then you can see if the Wooting keyboard sends any reports when you press keys (without using the analog sdk)

(on usage page 0xFF54 specifically)

[torpid fox]

I'll install wireshark again and try

[torpid fox]

yeah issue is definitely still happening with mouse 1 unplugged and mouse 2 in a different port

Gonna close the program now so I can use my computer normally

while installing

[torpid fox]

I can't find anything about a page

[quick bough]

Uhhh

Well, when you press a key, do you see a difference with and without Analog SDK?

[torpid fox]

There's definitely a difference and I don't have to press a key for it

when the SDK is running there is constant communication back and forth

[quick bough]

What's that communication say?

Or look like

[torpid fox]

It seems to be names of devices...?

As strings

[quiet root]

the sdk running or a program using the sdk

[torpid fox]

a program using the SDK

[quiet root]

well there will be coms when that runs as it would pull analog data from the keyboard constantly

unless its a test program that just inits and does nothing

[quick bough]

So is it like "the report being sent" is actually the OS constantly asking the HID if it has a report?

[torpid fox]

well since I wrote the program I can stop it from polling

see if that changes anything

[quick bough]

Okay, so according to GPT, the OS does have to poll the device for it to send any reports

Apparently USB 3.0 did introduce some event thing tho

[quiet root]

should always be present

[torpid fox]

I stopped it from reading the buffer or doing any get_analog calls and I still seem to have the mouse issue, FYI

[quick bough]

So, wireshark no longer shows the communication you were seeing before?

[torpid fox]

I haven't checked yet

[placid ledge]

mouse issues would probably be from it searching for devices, not from reading a device that's already opened

[torpid fox]

I opened wireshark 5 seconds ago

[quick bough]

Hmmm, what's the interval on the search for new devices?

[torpid fox]

That would explain a lot...

[placid ledge]

The firmware will always attempt to send a report when the analog data changes. My understanding is that the OS will not bother polling the interface if there are no readers of the interface, so that could be why you see less communication happening when the Analog SDK isn't running

[quiet root]

was about to say doesnt the sdk just read from the device when you wanna get the buffer?

[torpid fox]

I did see communication without the SDK. it was just only when I actually touched something instead of non-stop

[placid ledge]

yeah, it won't send multiple reports with the same information

so if you're not pressing anything, it will only have sent a report when the last pressed key was released, then it won't send anymore until you press a key

[torpid fox]

yeah

as expected

when the SDK is running there is constant back and forth though... mainly with device names?

[quick bough]

Not that I'm the best at reading Rust, but it looks like it scans for new devices every 500ms

[placid ledge]

That could be from the device scanning as well? As usually it won't need to get the strings as it would already have got them on device connection, but the device scanning may trigger it to get the strings again

[quiet root]

unsure if it might be nice to offload device search to SDK consumers

[placid ledge]

ideally it shouldn't be polling for devices, instead using OS events to check when a device connection happens

offloading the search to SDK consumers just moves the problematic baton to others 😂

[quick bough]

Maybe if the device scanning was changed from every 500ms to every 5000ms?

[placid ledge]

You could increase the period, and the affect would be minimized, but at the cost of having delayed detection for devices that are connected

Although, you could question how important that is to have

[quick bough]

hotswapping usb devices isn't exactly a hot path

[quiet root]

thats the point lol

[placid ledge]

ye but it doesn't solve the issue, it just makes the issue the responsibility of the consumers. Also increasing the work required (and potentially making breaking ABI changes) for the SDK consumers

[quiet root]

i mean how easy is it to consume USB events cross os

[placid ledge]

if any significant effort was to be put in to improving this behaviour, it would be the OS events path that I would go for

there's no unified approach, in general would be using specific apis

[quick bough]

depends on the device. for wooting devices is it very easy to use it on linux

but the linux apis regarding hid are poo

but "hidapi" abstracts away most of this pain

[torpid fox]

I just realized that the device scanning could explain why my hiccups happened less often with more mice connected

[placid ledge]

We've experimented with using this library: https://docs.rs/rusb/latest/rusb/ (libusb wrapper) to listen for device connect/disconnect events rather than polling

[torpid fox]

Since you're using past tense, did something go wrong while doing that and did you switch away?

[placid ledge]

Oh no, we're still using it. But the project is on hold and I haven't had much time to scrutinize it, but on the surface it appeared to behave pretty well. Although, I do remember it not working on all main OS' (Windows, Mac, Linux), I can't remember which ones worked/didn't work though.

[torpid fox]

If you want I can test it on all of them

[placid ledge]

Would be nice, although @remote junco may remember what the behaviour was like on each platform

[quick bough]

I don't experience any input lag by just constantly enumerating all HID devices on my machine

[placid ledge]

is that getAll function calling hidapi::refresh_devices?

[quick bough]

No, it's calling CM_Get_Device_Interface_ListW, HidD_GetAttributes, etc.

(I don't like external dependencies)

[torpid fox]

It's your soup again isn't it

[quick bough]

FWIW, it's very similar to hid_enumerate from hidapi

[quick bough]

It is my Soup again :^)

[placid ledge]

I wouldn't be surprised if the hidapi implementation causes more overhead than is necessary

[quick bough]

I don't even see a refresh_devices in hidapi

[placid ledge]

the development of the hidapi library was basically dead for many years until libusb picked it up fairly recently

Oh, I'm thinking of the hidapi-rs func, I guess it doesn't map directly onto the underlying library (it's been a bit since I used hidapi straight)

it calls hid_enumerate underneath

[quick bough]

So, it should be doing pretty much the same thing

I was originally doing a different way of enumerating HID devices, but then I stole the hidapi approach because I found it more elegant 😛

[placid ledge]

I think the lag coming from enumeration is fairly set-up specific. I've never really experienced it myself, but have seen a few different people run into it

I suspect it would depend a lot on your system's USB controllers/chipset, could also depend on the devices you have and how quickly they turnaround from string requests

[torpid fox]

There are a lot of string requests

[quick bough]

So, might be good to just not enumerate the devices as often as every 500ms?

[placid ledge]

I don't really get why it requests so many strings when they're the same ones that the OS already got when the device connected

[placid ledge]

500ms is quite a bit faster than it needs to be, but if polling can be eliminated, that's the real winner 😎

[quick bough]

It is possible that hidapi has overhead because it tries to populate all information when constructing the "hid_device" struct

things like product name and serial number I only lazily fetch on-demand

but I also don't see any input lag when spamming those requests

[torpid fox]

I'm going to check if the lag happens on my M2 Mac Mini; if it doesn't happen that means I get to work in peace (after seeing if I can help resolve or test anything with what's being discussed right now) and if it does happen that means the hardware setup should be mostly easily reproducible

[quick bough]

I would really like to know if you experience these lag issues with my evil C++ code, because I only scan devices until I find any analogue keyboard, then I just poll it for reports until its disconnected.

So, in theory, if device scanning is the issue...

[torpid fox]

Fine I'll try your plugin

but I'll restore my setup to a usable state first

[quick bough]

damn, a single AV vendor is foiling my plans for world domination

[torpid fox]

"AIDetectMalware"?

[quick bough]

Ah yeah, Bkav Pro detects literally every single C++ thingy lol

It even detects our fork of Lua

We've been trying to get their detections sorted

No response from them yet tho

[torpid fox]

Did a small bit of research and apparently bkav pro false positives aren't uncommon

Okay so your plugin detects my device

but other than that it does not work

[quick bough]

Oh?

[torpid fox]

I'll try seeing if calling read_analog instead of read_full_buffer does anything

oh actually the whole godot editor doesn't respond now

oh it just returned

weird

[quick bough]

Btw., you wouldn't know if it detected your keyboard or not

it just provides a dummy to the SDK

because I'm lazy

[torpid fox]

oh

[quick bough]

If you don't mind, here's a standalone analogue keyboard visualisation that would say if it detected your keyboard or not.

[torpid fox]

I just got this helpful error I wrote a while ago

[torpid fox]

uhhh no I'mma pass

[quick bough]

LOL

We love heuristics

Here's the same EXE but in a debug configuration; now it has only 1 detection again 🙂

It is basically just this: https://github.com/calamity-inc/Soup/blob/senpai/CLI/cli_keyboard.cpp

[torpid fox]

I mean if I were to assume that 10 detections one was malware and this is that same thing but in the form of a debug build
It just has a hash that hasn't been seen before meaning it might as well be malware

[quick bough]

Well, neither hash has ever been seen before

It's just AI heuristics rolling the dice

😄

[torpid fox]

idk bitdefender says it's gibon

[quick bough]

I've submitted a false-positive report with bitdefender

[torpid fox]

well I am running the program

the visuals work

[quick bough]

Usually I don't bother, this is software we deploy to 100k+ users 😄

AV industry definitely not being a joke 😄

[torpid fox]

anyway

I'll try switching to your plugin again because I just realized I disabled it

[quick bough]

Anyway, yeah, it is weird if the program would work but not the plugin

[torpid fox]

(I swapped it out because I was planning on using stock again)

[quick bough]

It should look like this

And the other plugins are just stashed away here

[torpid fox]

I know

I know what I did and I did it intentionally

[quick bough]

Alright, cool

[torpid fox]

still works with your plugin

[quick bough]

The soup.exe is standalone and does not rely on the Wooting Analog SDK at all

It polls your HID devices directly

[torpid fox]

okay

well it works

[quick bough]

Without mouse input lag?

[torpid fox]

Haven't noticed any so far no

[quick bough]

I am going to compile this

If it was the device enumeration, this should lag your mouse a lot

[torpid fox]

Yep I was able to gather that idea from the code

given what I saw in wireshark

YEP

Mouse is literally unusable

can't move it at all

[quick bough]

Interesting

I am going to remove the polls for product name & serial number

[torpid fox]

mouse is fine

I think we found the culprit

[quick bough]

Cool, so what Wooting needs now is a custom fork of hidapi that doesn't poll the product name and SN of every device every 500ms haha

Not so useless that I reinvent every wheel after all

[torpid fox]

@placid ledge apologies for the ping but I think we found something here

Something you might want to know

[quick bough]

Unrelated but did you manage to get my plugin working or that still doesn't work with your app?

Also, the relevant code that was causing the lag:

A workaround might be to call hid_enumerate using only the Wooting Vendor IDs

because then hidapi won't try to initialise the hid_device for anything that doesn't match

[torpid fox]

still completely nonfunctional

[quick bough]

Welp, that's strange

Your device doesn't happen to be the Wooting UwU, does it? xD

That's the only one the 0.1.0 version doesn't support

[torpid fox]

Nah it's a Two HE ARM

[quick bough]

Welp, maybe if you ever publish your app, I can see if it works for me

[torpid fox]

I was thinking of doing that when my to-do list for it is done but I guess I can do it now

I should clean up a bit first though

[quick bough]

or maybe I should ask if it works for you with the "Woot-verlay"?

because that did work for me, although not brilliantly, still need to make slight improvements

[torpid fox]

Works with woot-verlay

[quick bough]

Okay, then something you do must be weird, I'd definitely love to take a look

[torpid fox]

It could be because I am using wooting-analog-wrapper?

[quick bough]

idk, tbh. in theory it should "just work"

[torpid fox]

yeah

[torpid fox]

Note that you should really use cargo and that I don't think you'll be able to rustc your way out of this one

[quick bough]

I don't write any Rust code?

[torpid fox]

thing is though, to run my project you currently have to build it yourself

unless I give you a build

in which case I might as well provide a whole godot project

[quick bough]

uhhh

Well, I don't have godot either lol

[torpid fox]

uhh good luck then?

[quick bough]

I shall try using the wooting-analog-wrapper cargo crate

[torpid fox]

Sure.

[quick bough]

FWIW, I have used cargo for a few rust test projects in the meanwhile

although all of those were dreadful

[torpid fox]

it's not on crates.io. get it from the wooting analog repo by adding ```toml
wooting-analog-wrapper = { git = "https://github.com/WootingKb/wooting-analog-sdk", branch = "develop" }

this is really unconventional and I don't like it either

[quick bough]

Looks like something didn't go so well

[torpid fox]

that's a very big malloc

did you try to read_full_buffer with too big of a number?

[quick bough]

I just copy-pasted this: https://github.com/WootingKb/wooting-analog-sdk/blob/develop/wooting-analog-wrapper/src/bin/main.rs

[torpid fox]

I mean I don't think you're supposed to do that but it seems like that isn't the issue here

[quick bough]

That's odd, this looks like an address

Well, it's definitely Rust code doing this

Doesn't seem like it provides a backtrace for this bogus memory allocation either

[torpid fox]

there's a command line argument you use if you want a backtrace usually

[quick bough]

I did set the RUST_BACKTRACE environment variable to "on"

[torpid fox]

oh right it was an environment variable not an argument

my bad

[torpid fox]

tell me how you're so sure

[quick bough]

I basically removed all the code in my DLL and it still crashed

[torpid fox]

but does it crash when it is working with the official wooting plugin

[quick bough]

Yeah, no.

It might just be an issue with C plugins tho

[torpid fox]

are you saying it crashes or not

[quick bough]

It doesn't

[balmy iron]

How about running in a debugger, do you see where it's failing then?

[quick bough]

Will try

debugger doesn't seem to catch anything

[balmy iron]

Lovely

[quick bough]

nevermind attached to wrong program lol

Yeah, it is entirely in the Rust code here

I don't have a PDB for this, so I don't know what's exactly happening

[torpid fox]

If only you could build it yourself

(I have a pdb)

(you could get one too if you build it)

[quick bough]

Well, I think it's an SDK issue, I raised it here: https://github.com/WootingKb/wooting-analog-sdk/issues/59

But if I don't use get_connected_devices_info, my plugin seems to work fine with wooting-analog-wrapper. ¯_(ツ)_/¯

[torpid fox]

Except I don't use get_connected_devices_info at all

I did write bindings for it, it's just dead code rn

not being used

[quick bough]

Hmm, well, then it also works fine for me using this barebones driver for wooting-analog-wrapper

[floral crane]

I found the ETA but it's late, someone should probably change that on the website. it's a week later and stable still doesn't support the UwU

Is this all you need for udev rules now? you don't need all the gamepad and xinput stuff for xboxdrv anymore? https://help.wooting.io/en/article/wootility-configuring-device-access-for-wootility-under-linux-udev-rules-r6lb2o

[quiet root]

didnt need xbox stuff for ages

someone put it into xboxdrv or smth

[floral crane]

This one still says all that stuff but it's not linked anywhere, I had to search for it, and it doesn't have newer devices. https://help.wooting.io/en/article/guide-configuring-xinput-support-for-linux-69m32u

Oh I see, it was put into the linux kernel https://github.com/torvalds/linux/blob/master/drivers/input/joystick/xpad.c#L364-L369

someone should put the tkl on there before it releases so there's a chance it's in the kernel steamos uses by the time it comes out

[quiet root]

see #🤖│community_dev message

although 2he arm and uwus are missing in that

[floral crane]

backports are amazing

uwu can't be a gamepad can it?

[quiet root]

it can

it has all the features of our normal keyboards

[floral crane]

hmm 3 buttons, I guess it'll work

[quiet root]

just less keys

i suspect people would need to submit another patch to the kernel though so it can use the new devices

sad we cant list the entire vendor id

[floral crane]

pre-list the next 10 devices unnamed 🤔

updating the aur packages now

https://aur.archlinux.org/packages/wootility-lekker-beta-appimage
I'll update stable when it releases

[quick bough]

I had to do that udev rule stuff for my Wooting Two HE

[quiet root]

the udev rules are still necessary yes but not special xbox setup anymore

[strong siren]

anyone with a uwu rgb willing to test Artemis / RGB.NET? should work as long as you have a dotnet runtime working, windows or linux is fine

[quiet root]

macos?

[strong siren]

technically no explicit reason it shouldnt work but i've never tested it

even less so on ARM

[quiet root]

well last i tried it crashed immediately

[strong siren]

did you test artemis or the console test app

[quiet root]

artemis

[strong siren]

test app would be a good start to see how dotnet handles dllimports on arm

do you have a dotnet dev env setup or should i make a binary?

based off this https://github.com/diogotr7/RGB.NET/tree/sandbox

[quiet root]

i have no dotnet setup on mac

[strong siren]

i have no idea if any of those will work but there you go

these are published as self contained so shouldnt need runtime installed or anything

[quiet root]

never used rgbnet directly

[strong siren]

thats just a console app that should render a rainbow moving left to right on all devices it finds

ive tested it pretty extensively on two he + one + 60he

[quiet root]

just dies immediately

assuming i need to run RGB.NET.Sandbox

[strong siren]

yes

[quiet root]

dw about the rename but macos doesnt like . in executables

[strong siren]

any diff with the x64 one?

[quiet root]

yeah that one does this

[strong siren]

that's something

i assume with this running as x64 you need the sdk binaries to be x64 as well? are the binaries in mac releases x64 or arm?

[quiet root]

our rgb sdk? x64

[strong siren]

i added some extra logging to this one but apart from that i have no idea whats wrong

gotta look into making a mac vm or something lol

[quiet root]

wait but the rgb sdk doesnt support the uwu yet

[strong siren]

theres a branch

i asked simon and he told me it should work fine

but yeah not that dll/dylib

[quiet root]

i checked the dylib is in there

no dice

[strong siren]

hm ok thanks for the testing

im not suire how much of this is mac or arm issues

[quiet root]

might be some sort of permission thing

[strong siren]

i just got a macos vm setup on my unraid box so ill try to fix things

the error code 0 is the most stupid part, no info whatsoever

[quick bough]

Silly question, but in this Razer press release from 2019, Wooting is mentioned as one of the brands that apparently supports Razer Chroma RGB, yet my Wooting keyboard doesn't show up in Razer Synapse/Chroma.

[strong siren]

if you install wootility it has an integration

it's pretty shitty though, use mine :)

[quick bough]

What's yours?

[strong siren]

https://github.com/diogotr7/RazerSdkReader

or use artemis which uses this library

that repo has a tray app that targets wooting keyboards specifically with razer chroma games

[quick bough]

Well, that is sorta what I am trying to achieve, yeah

[quiet root]

chroma connect is a joke

[strong siren]

it is

[quick bough]

I was gonna replace the RzChromaSDK64.dll and just redirect the API calls

but also I was not trying to get banned from some games :^)

[strong siren]

we used to do that back in the day

for aurora

i have dlls that do this and write the colors to a pipe

https://github.com/Artemis-RGB/Artemis.Plugins.Wrappers

but it's not worth using because bans and also some (most) games sigcheck the dll before loading it

[quiet root]

the solution is that USB should finally add RGB into their standard so windows and other OSs can make a standard way to interact with RGB devices

[strong siren]

LampArray is basically this as far as i understand it

[quick bough]

I will try installing the Wootility first, but your project looks good as well

[strong siren]

it's limited to 5 colors

[quiet root]

imma switch to win and then i can try artemis on uwus

[strong siren]

the razer one. your keyboard is split into 5 zones

[quiet root]

got an artemis build artifact for uwu support?

[quick bough]

Patiently waiting for my UwU to arive >.<

[strong siren]

no, you'll have to shuffle dlls around for now sadly :p

[quiet root]

send dlls

[quick bough]

Woaw, hello there

[strong siren]

yeah thats the one

[quiet root]

as mentioned keep in mind 5 zones only

[quick bough]

wdym 5 zones only?

[quiet root]

its not individual keys

[strong siren]

it gets your keyboard, spits it into 5 zones horizontally, and colors each of those

[quiet root]

just 5 blobs of color basically

[quiet root]

thats why razer chroma for third party keyboards is a joke

[strong siren]

it's equivalent to that "chromalink' window on the right

[quick bough]

?!

[strong siren]

with mine you get the rest like the keyboard on top

[quick bough]

Welp, I'm getting colour at least now!

Will see about the granularity

Yeah okay this is horrible

[strong siren]

i can probably clean up the wooting tray app and make an installer etc if you think people care about it

[quiet root]

exactly why i said its a joke

its razers poor attempt at seeming nice and inclusive

[quick bough]

I don't think they do, I just noticed the RGB while trying the Razer Huntsman V2 Analog for a few days

and I decided I want it on the wooting keyboard now as well

I'm gonna plug it in again and compare the RGB effects

[strong siren]

gimme a minute ill ready it up

still waiting for this smh

[quiet root]

thats only half the wait

after that it needs to do the OOBE setup lol

[strong siren]

ah thats great

[quiet root]

although hmm

10.9

quite old

might be rather quick on oobe

[strong siren]

yeah i know it was the first one i found lol

[quiet root]

prob also one of the last ones that work on non arm

hackintosh will have a hard time soonish when intel support drops from macos

[quick bough]

Why is it so bad with the official integration

[strong siren]

because razer

[quiet root]

also hope you have success in actually running the macos vm as some pc setups just make the vm freeze

[strong siren]

i was gonna put this in the artemis discord but you left

[quick bough]

So, how do I use your thing dio, do I have to build it myself in VS?

[quiet root]

oh yeah i stopped using artemis long long ago

[strong siren]

if you have a dev env setup yeah that's the easiest

actually ill make a CI script for it

[quick bough]

Why did you not write this in C++ tho >.<

[strong siren]

because i hate c++

[quick bough]

why

Also, slight problem: The top row is entirely black.

[strong siren]

what kb?

[quick bough]

Two HE

[strong siren]

maybe i broke it lemme test

what game are you testing with?

[quick bough]

Warframe

Avalonia seems to be showing it correctly

[quiet root]

gg artemis crashes with woot plugin

[strong siren]

pog

does it crash with any info

[quiet root]

nope

just closes

[strong siren]

can you please check event viewer

[quiet root]

CoreCLR Version: 7.0.1123.42427
.NET Version: 7.0.11
Description: The process was terminated due to an unhandled exception.
Exception Info: System.AccessViolationException: Attempted to read or write protected memory. This is often an indication that other memory is corrupt.
Stack:
   at WootingAnalogSDKNET.WootingAnalogSDK.initialise()
   at WootingAnalogSDKNET.WootingAnalogSDK.initialise()
   at WootingAnalogSDKNET.WootingAnalogSDK.Initialise()
   at Artemis.Plugins.Devices.Wooting.Services.AnalogService.WootingAnalogService.Activate()
   at Artemis.Plugins.Devices.Wooting.Services.ReusableService.ActivateInternal()
   at Artemis.Plugins.Devices.Wooting.Services.ReusableService.RegisterUse()
   at Artemis.Plugins.Devices.Wooting.Modules.WootingAnalogModule.Enable()
   at Artemis.Core.PluginFeature.InternalEnable()
   at Artemis.Core.Modules.Module.InternalEnable()
   at Artemis.Core.Modules.Module`1[[System.__Canon, System.Private.CoreLib, Version=7.0.0.0, Culture=neutral, PublicKeyToken=7cec85d7bea7798e]].InternalEnable()
   at System.Threading.ExecutionContext.RunFromThreadPoolDispatchLoop(System.Threading.Thread, System.Threading.ExecutionContext, System.Threading.ContextCallback, System.Object)
   at System.Threading.Tasks.Task.ExecuteWithThreadLocal(System.Threading.Tasks.Task ByRef, System.Threading.Thread)
   at System.Threading.ThreadPoolWorkQueue.Dispatch()
   at System.Threading.PortableThreadPool+WorkerThread.WorkerThreadStart()

[strong siren]

hmm analog

if you unplug the uwu does it still crash

[quiet root]

i didnt have it plugged in yet

only 2he arm plugged in

[strong siren]

i have two he (avr i think?) and 60he arm rn

working well

[quiet root]

i replaced the plugin again to be sure still crashing

[strong siren]

odd, do you have the analog sdk installed? maybe it crashes if not installed. i recently changed that lol

[quiet root]

yes even updated

[strong siren]

i can confirm that the analog features work in artemis with sdk version 0.7.2

[quiet root]

0.7.4 is current

and wootility will install that

[strong siren]

no that's not the issue, just updated

still good

[quiet root]

odd then

[strong siren]

yeah working fine with both kbs

[quiet root]

whny are the sdks in the zip from 07

did the branch exist in july?

[strong siren]

im an idiot lol, but that wont fix this analog thing

ill update them

[quiet root]

updated analog wrapper now it doesnt close but shows a fail

[strong siren]

analog wrapper?

[quiet root]

its in the folder

le content

[strong siren]

ah the binary

yes

will do as wel

[quiet root]

 ---> Artemis.Core.ArtemisPluginException: Failed to initialise WootingAnalog SDK: Failure
   at Artemis.Plugins.Devices.Wooting.Services.AnalogService.WootingAnalogService.Activate() in D:\Source\Artemis\Artemis.Plugins\src\Devices\Artemis.Plugins.Devices.Wooting\Services\AnalogService\WootingAnalogService.cs:line 80
   at Artemis.Plugins.Devices.Wooting.Services.ReusableService.ActivateInternal() in D:\Source\Artemis\Artemis.Plugins\src\Devices\Artemis.Plugins.Devices.Wooting\Services\ReusableService.cs:line 40
   at Artemis.Plugins.Devices.Wooting.Services.ReusableService.RegisterUse() in D:\Source\Artemis\Artemis.Plugins\src\Devices\Artemis.Plugins.Devices.Wooting\Services\ReusableService.cs:line 17
   at Artemis.Plugins.Devices.Wooting.Modules.WootingAnalogModule.Enable() in D:\Source\Artemis\Artemis.Plugins\src\Devices\Artemis.Plugins.Devices.Wooting\Modules\WootingAnalogModule.cs:line 24
   at Artemis.Core.PluginFeature.InternalEnable() in D:\a\Artemis\Artemis\Artemis\src\Artemis.Core\Plugins\PluginFeature.cs:line 178
   at Artemis.Core.Modules.Module.InternalEnable() in D:\a\Artemis\Artemis\Artemis\src\Artemis.Core\Plugins\Modules\Module.cs:line 317
   at Artemis.Core.Modules.Module`1.InternalEnable() in D:\a\Artemis\Artemis\Artemis\src\Artemis.Core\Plugins\Modules\Module.cs:line 85
   at System.Threading.ExecutionContext.RunFromThreadPoolDispatchLoop(Thread threadPoolThread, ExecutionContext executionContext, ContextCallback callback, Object state)
--- End of stack trace from previous location ---
   at System.Threading.ExecutionContext.RunFromThreadPoolDispatchLoop(Thread threadPoolThread, ExecutionContext executionContext, ContextCallback callback, Object state)
   at System.Threading.Tasks.Task.ExecuteWithThreadLocal(Task& currentTaskSlot, Thread threadPoolThread)
   --- End of inner exception stack trace ---
   at System.Threading.Tasks.Task.ThrowIfExceptional(Boolean includeTaskCanceledExceptions)
   at System.Threading.Tasks.Task.Wait(Int32 millisecondsTimeout, CancellationToken cancellationToken)
   at System.Threading.Tasks.Task.Wait(TimeSpan timeout, CancellationToken cancellationToken)
   at System.Threading.Tasks.Task.Wait(TimeSpan timeout)
   at Artemis.Core.PluginFeature.SetEnabled(Boolean enable, Boolean isAutoEnable) in D:\a\Artemis\Artemis\Artemis\src\Artemis.Core\Plugins\PluginFeature.cs:line 151
   at Artemis.Core.Services.PluginManagementService.EnablePluginFeature(PluginFeature pluginFeature, Boolean saveState, Boolean isAutoEnable) in D:\a\Artemis\Artemis\Artemis\src\Artemis.Core\Services\PluginManagementService.cs:line 701
   at Artemis.Core.Services.PluginManagementService.EnablePlugin(Plugin plugin, Boolean saveState, Boolean ignorePluginLock) in D:\a\Artemis\Artemis\Artemis\src\Artemis.Core\Services\PluginManagementService.cs:line 516
   at Artemis.UI.Screens.Plugins.PluginViewModel.<UpdateEnabled>b__53_4() in D:\a\Artemis\Artemis\Artemis\src\Artemis.UI\Screens\Plugins\PluginViewModel.cs:line 162
   at System.Threading.ExecutionContext.RunFromThreadPoolDispatchLoop(Thread threadPoolThread, ExecutionContext executionContext, ContextCallback callback, Object state)
--- End of stack trace from previous location ---
   at System.Threading.ExecutionContext.RunFromThreadPoolDispatchLoop(Thread threadPoolThread, ExecutionContext executionContext, ContextCallback callback, Object state)
   at System.Threading.Tasks.Task.ExecuteWithThreadLocal(Task& currentTaskSlot, Thread threadPoolThread)
--- End of stack trace from previous location ---
   at Artemis.UI.Screens.Plugins.PluginViewModel.UpdateEnabled(Boolean enable) in D:\a\Artemis\Artemis\Artemis\src\Artemis.UI\Screens\Plugins\PluginViewModel.cs:line 162```

now i get this

[strong siren]

yeah that's prob fine

gimme a minute

or build the rgb sdk binary yourself if you have it handy

[quiet root]

i dont think i have a c build env setup on windows

[strong siren]

i only updated the windows rgb binary and the analog wrapper

[quiet root]

 ---> System.MissingMethodException: Method not found: 'Void Artemis.Core.Services.IDeviceService.AddDeviceProvider(Artemis.Core.DeviceProviders.DeviceProvider)'.
   at Artemis.Plugins.Devices.Wooting.WootingDeviceProvider.Enable()
   at Artemis.Core.PluginFeature.InternalEnable() in D:\a\Artemis\Artemis\Artemis\src\Artemis.Core\Plugins\PluginFeature.cs:line 178
   at System.Threading.ExecutionContext.RunFromThreadPoolDispatchLoop(Thread threadPoolThread, ExecutionContext executionContext, ContextCallback callback, Object state)
--- End of stack trace from previous location ---
   at System.Threading.ExecutionContext.RunFromThreadPoolDispatchLoop(Thread threadPoolThread, ExecutionContext executionContext, ContextCallback callback, Object state)
   at System.Threading.Tasks.Task.ExecuteWithThreadLocal(Task& currentTaskSlot, Thread threadPoolThread)
   --- End of inner exception stack trace ---
   at System.Threading.Tasks.Task.ThrowIfExceptional(Boolean includeTaskCanceledExceptions)
   at System.Threading.Tasks.Task.Wait(Int32 millisecondsTimeout, CancellationToken cancellationToken)
   at System.Threading.Tasks.Task.Wait(TimeSpan timeout, CancellationToken cancellationToken)
   at System.Threading.Tasks.Task.Wait(TimeSpan timeout)
   at Artemis.Core.PluginFeature.SetEnabled(Boolean enable, Boolean isAutoEnable) in D:\a\Artemis\Artemis\Artemis\src\Artemis.Core\Plugins\PluginFeature.cs:line 151
   at Artemis.Core.Services.PluginManagementService.EnablePluginFeature(PluginFeature pluginFeature, Boolean saveState, Boolean isAutoEnable) in D:\a\Artemis\Artemis\Artemis\src\Artemis.Core\Services\PluginManagementService.cs:line 701
   at Artemis.UI.Screens.Plugins.Features.PluginFeatureViewModel.<UpdateEnabled>b__47_0() in D:\a\Artemis\Artemis\Artemis\src\Artemis.UI\Screens\Plugins\Features\PluginFeatureViewModel.cs:line 207
   at Avalonia.Threading.DispatcherOperation.InvokeCore()
--- End of stack trace from previous location ---
   at Artemis.UI.Screens.Plugins.Features.PluginFeatureViewModel.UpdateEnabled(Boolean enable) in D:\a\Artemis\Artemis\Artemis\src\Artemis.UI\Screens\Plugins\Features\PluginFeatureViewModel.cs:line 207```

thats the device provider

[strong siren]

bruh, yeah im working off of artemis dev and not master, 1 sec

sorry and thank you for the patience :p