🤖│community_dev

[sweet peak]

sounds like a rights issue

[quick bough]

Should we bring it to the UN, or?

[quick bough]

Welp, I just enabled analogue reports for the Huntsman Mini and it's not working anymore, so I think the openrazer driver just doesn't have the same analogue input support as the official Windows software

[sweet peak]

human rights might be a bit too far

but we'll get it to Den Haag

[quick bough]

Damn, I was just there recently

and Simon wasn't even there so I couldn't bully him

[quick bough]

Wait, he released an RGB SDK update yesterday, but still no analogue SDK update

[quick bough]

A bit less interesting but I just got around to porting digital keyboard visualisation to Linux

It's a bit cringe because I have to parse the HID report descriptor and then read the input reports myself, but that's just the Linux way I guess

[quick bough]

https://analogsense.org/Switchy/

[quick bough]

Welp, still can't successfully flash firmware on Firefox (it would be nice if one of you could maybe increase the fail timeout), but at least importing & applying profiles now seems to work without it ever timing out.

Not entirely sure why it times out at all, but Wootility seems incredibly timing-sensitive.

Oh wait, that's interesting, it seems like maybe one of the responses was dropped

That was a stupid bug, but I think Wootility should now work flawlessly in Firefox... if we ignore the gamepad buttons being way too big.

[sweet peak]

im impressed

[grizzled wolf]

Hi, I'm trying to use this rust library to play with rgb on my 80HE but I noticed some of the keys are not mapped correctly e.g. Key::F1 lights up F2 I wonder if this is something that's known(firmware issue) / my keyboard being weird / wooting-rgb-sdk issue
https://github.com/ShayBox/Wooting-RGB

[quick bough]

I'd say open an issue with that package, since the way the RGB SDK works is by (x, y) and not based on keys. It seems the crate assumes that F1 is at (0, 2), but the 80HE has it at (0, 1) because of space efficiency.

[grizzled wolf]

tried changing the touple for F1 to point at (0,1) my F1 key now lights up

[quick bough]

@floral crane

[grizzled wolf]

I find it weird that they have a picture of the keyboard matrix for "keyboards" but it doesn't apply to 80HE

[quick bough]

Eh, yeah, but it's meant to be used for effects, not for specific keys, I guess

and that picture is clearly for a 100% keyboard like the Wooting Two, I don't think it can be moddeled to be the same on the 80HE

[grizzled wolf]

I wonder if it works for 60HE

[quick bough]

I unfortunately don't own a 60HE so I don't know

[strong siren]

the matrix hasnt been updated yet

here are correct mappings in the meantime, we'll update the matrix in the readme soon.

https://github.com/DarthAffe/RGB.NET/blob/a26eaf9a9aad71f035a89752c81422b60bd8e59d/RGB.NET.Devices.Wooting/Generic/WootingLedMappings.cs#L114

[quick bough]

That 60% mapping is just cursed

[grizzled wolf]

this one also isn't completely correct and is missing some leds
Keyboard_Home should be Page Up with the default configuration
Keyboard_Home should be Page Down with the default configuration
Space key now has 5 leds at (5, 4), (5, 5), (5, 6) <- the key, (5, 7), (5, 8)

[floral crane]

Afaik the matrix has never changed and it's the same for all devices except for missing keys, but I've never touched rgb with the sdk or the rgb crate, I only use commands with the sys crate.

[quick bough]

That's obviously not correct since on the 80HE, the F1 key is (0, 1) instead of (0, 2)

[floral crane]

F1 should be 0,1 anyway, that's what the matrix pic has always shown

[quick bough]

No?

[floral crane]

wait what why is there a gap

[quick bough]

because there is a gap

just not on the 80HE

[floral crane]

and why is it not there for the 80HE? the rgb sdk internally has no different matrix for the 80he

[quick bough]

what keyboard are you typing on right now?

[floral crane]

lekker

[quick bough]

So, Wooting Two LE?

[floral crane]

yeah

[quick bough]

Then look at that area on your keyboard

there's a very clear gap between esc & F1

the exact same space that's above the '1' key

It's just a lot tighter on the 80HE

so there is no such space

[strong siren]

the mappings i linked above definitely light up every key, just some ids might be wrong apparently :p

[floral crane]

Is the 80HE the first keyboard with a different matrix?

The matrices should be defined in the rgb sdk source for programmatic use, instead of having everyone hard-code their own based on a picture, and swapping it based on the device model themselves

I wonder if you could make a float aligned matrix for precisely aligned key positions, instead of a grid aligned matrix where keys are not aligned, like enter, shift, plus, and every key in a grid instead of shifted. Probably a big change though

[strong siren]

that would be nice for lamparray

[quick bough]

that would be nice if the firmware provided a LampArray interface 😛

it's kind of funny, tho. the analogue SDK only provides key codes, and the RGB SDK only takes positions.

there are no compromises on that and no way to reliably convert between the two

like, have fun determining what the "correct" layout of the Wooting UwU is

[balmy iron]

Yeah would be nice if the devices just used LampArray directly but that might be too much work, since it works with physical lamp positions

And also has special markings for "this is for key K" and whatnot

[quick bough]

I guess your best bet is just taking heavy "inspiration" from the Wootility's beautiful source code and just figuring out some way to map keycode to RGB position with that

and then weep once you realise that a key can be present more than once

(same issue with trying to map the inverse for analogue input, except with RGB you can at least be pragmatic and just light up all matching keys)

[quick bough]

@placid ledge You might want to hold back 9e8e3b16a383e870ac85bd1b1b56ac2a7dbefdab[1] until a later release. This is likely causing the CI to fail.

[1] https://github.com/WootingKb/wooting-analog-sdk/commit/9e8e3b16a383e870ac85bd1b1b56ac2a7dbefdab

[placid ledge]

Thanks for the heads up. I don't think it's why the CI is failing right now (it seemingly isn't even executing the cplugin for some reason).

But it likely does explain why it's failing for me locally

Ugh, analog sdk project is such a mess, I really gotta find some time to properly get it into shape

[quick bough]

Rewrite in C++ 🙏

Well, you are right, the CI is failing even with that reverted

but I just figured since it's ABI breaking

[grizzled wolf]

Please don't 🙏

[balmy iron]

Please do, it would be funny

[quick bough]

Goddess forbid it actually be usable or maintainable

but nah, I think the analogue SDK is fine as it is, it's just a very niche kinda thing. very few people are interested in implementing it, and even less are willing to stick with it after they have any sort of issue with it.

the only thing you can do is hope to reduce the number of issues people will have with it, which I think means not trying to make everything an abstract problem that needs abstract solutions

which is how you end up with splitting everything over multiple DLLs and then running into crashes because Rust has an unstable ABI 😄

another issue with that approach is that 32-bit processes can't load 64-bit DLLs but I suppose it's a moot point given that 32-bit is basically dead nowadays

but I also can't think of a better way to do it that maintains the same kinda extensibility with plugins

Well, okay, you could always just ship a Lua VM and make HID bindings for Lua 😄

[quick bough]

Funny, apparently you guys also omitted the .dll.lib back when it was still written in C

You guys need to do another CODE IT contest, tho 😛

[umbral carbon]

Heyllo! I need some help with a ZMK port to an RP2040-based board. My config files look decent, but I keep getting held up at the following error:

(in my CI)

devicetree error: /tmp/zmk-config/zmk/app/boards/seeeduino_xiao_rp2040.overlay:7 (column 1): parse error: expected ';', not 'xiao_serial'

[quick bough]

It might help to provide the source code and not just error logs: https://github.com/pmnlla/zmk-board-cfg

Unfortunately I don't even know what you are doing. I only know QMK.

[umbral carbon]

That's fair. Source is that, yes. ambrylight branch is to be updated soon when I get that working.

I've also finished more recent PR runs, solved some issues, and ran into another one I can't quite figure out. Will update tomorrow.

[placid ledge]

When I get some more time to spend on the SDK, I want to just kill the whole rust plugins and have it be included in the core bit and just have the C ABI for plugins.

The biggest mistake I made in the architecture of it was trying to make it too flexible, it's far too split up which just creates a whole lot of headaches

[quick bough]

Yeah. I was trying to do the same thing, then I learned Rust doesn't have static initialisers.

Regarding the old analogue SDK, tho, is there no compatibility shim?

Like, the DLL I have here hardcodes the FF01 PID, and I'm not even sure my original Wooting Two would work with this because I probably updated the firmware at some point

[placid ledge]

You can do static initialisers, before you had to do it with lazy_static, but they have a built-in replacement for that now

[placid ledge]

There isn't, I don't exactly remember why we never did that

[quick bough]

Well, I might be about to do that

[placid ledge]

[quick bough]

but first gotta check my Wooting Two, I guess

Yup, RIP me, I updated the firmware

I figure there's no way to downgrade it either

[placid ledge]

You can technically downgrade it, but it will corrupt the serial number on-board

Although, I'm not sure downgrading it would necessarily help you try out the old sdk. IIRC the analog interface was altered for the "new" analog sdk, even on v3 Wootility/v1 firmware

[quick bough]

isn't there funny command I can send to change the serial

at least I saw some related code in wootility

[quick bough]

Well, yeah, I want to downgrade it so it's compatible with that old analogue SDK

[placid ledge]

ye, the report structure that the old sdk expects is very different and you won't be able to get that by downgrading to the latest v1 firmware

[quick bough]

damn

Wait, where is the original Wooting Two here?

[placid ledge]

That's also prob why we never created a shim, cus the old sdk works on the basis of the row/col / scancode. Whereas the "new" one works on hid codes. So I'm not sure you can really create a good shim

[quick bough]

Of course you can, just mappy map it

[placid ledge]

The "Wooting Two (HE)" is a catchall for Wooting Two's

Wait, why is Lekker edition there, it's supposed to be grouped Was thinking of v5, it's a full "Wooting Two" catchall

[placid ledge]

You can make an okay shim that way, not a good one

[quick bough]

Of course it makes an assumption that keys are mapped as expected

[placid ledge]

ye, but what even is the motivation for making a shim anyway?

[quick bough]

Make some old code that was made for the CODE IT contest work

16 and 18 KB for the original SDKs tho is impressive

I guess I'll just make the shim and que sera sera

Oh, there it is, in the wrapper\sdk directory. How silly of me not to check there.

[quick bough]

Welp, could not get the thingy working

Oh, I am so cooked, this is 32-bit

[quick bough]

Got it working: https://github.com/AnalogSense/old-wooting-analog-sdk-shim

Obviously can't use the "new" analogue SDK for 32-bit, but oh well

On another note, are there any scancodes except 0XX, 3XX, and 4XX? Because I'm thinking with some aggressive bit packing, both the scancode and the key position could fit in 2 bytes.

[quick bough]

Or rather, if there are less than 512 total scancodes (which I suspect there are), then it is possible to pack all of them in 9 bits, leaving 7 bits for the position, which can be encoded as y * 21 + x to make it fit.

Another assumption one can make is that 4XX is actually only 0x400..0x41F and that 3XX is actually 0x320..0x3FF and then you get this rather "unintrusive" way of encoding them:

[sweet peak]

On https://github.com/WootingKb/wooting-analog-sdk?tab=readme-ov-file

Follow the installation instructions for the Wooting Analog Plugin
That has a link that points to a repo that says it has been merged into the wooting analog sdk

[strange kite]

would it be possible to modify the tickrate on my 80HE?

Essentially I speedrun a game that runs at 31fps and I want the shortest possible inputs to be 32ms

[quick bough]

Uhh, no, I don't think there's a feature like that

[strange kite]

would you mind explaining how the community dev thing works, like can people code their wootings or somthin?

[quick bough]

Well, it's mostly about analogue input and RGB output

(Actually it's mostly coping with Rust)

[strange kite]

so you cant upload your own code?

[quick bough]

Nope

You could try Keychron's Q HE series of keyboards. Their firmware is somewhat open-source.

[strange kite]

because its already a feature i just want to modify the time

[strange kite]

yeah ik you can do it on there but i assumed a $300 wooting could do the same

[quick bough]

lol, why would they open-source their juicy firmware

it's so low latency and shit

big industry secrets!!!

[quick bough]

Where even is this setting

[strange kite]

under the advanced tab in dks and mod tap

[quick bough]

Well, bad news, looks like the options are hard-coded

[strange kite]

what does that mean?

[quick bough]

It means you're at Simon's mercy

Godspeed 🙏

[strange kite]

im sorry?

[strange kite]

where did you find the code?

[quick bough]

Just looking at what Wootility sends to the keyboard

[quick bough]

You should probably put your "feedback" in an appropriate place, e.g. either here in the #1141096226884956181 channel, or on Github: https://github.com/WootingKb/wootility-issues/issues

[strange kite]

how are you viewing that?

[strange kite]

ok will do thanks

[quick bough]

I'm on Firefox, so I had to write my own WebHID provider, so I can just insert logging there

[strange kite]

oh wow ok thanks for the help

[quick bough]

Does "NKRO Disable" just change the input report format to use "array" instead of "variable"? If so, would be weird if there are actually devices with such an implementation hole. Also, why only 6KRO? Surely an array of size 16, or 80 would also be fine?

[shy vale]

sounds like you're trying to eliminate 0-framers?

i'm not sure if you're looking to articifially extend inputs shorter than 32ms to be 32ms long, or you're looking to drop inputs shorter than 32ms altogether

[quick bough]

he wants to extend them to be at least 32ms

I think the PR was fine last I checked, but I guess I'll just double-check if my plugin still works with ABI version 1

Well, somehow it just panicked

[placid ledge]

6KRO is mainly for compatibility, so it's in the most common format. Not really much point in pushing it hard when anything that will properly parse HID reports should be able to handle the full nKRO interface

[placid ledge]

Seemingly it's with the DeviceInfo conversion? Are you fully using DeviceInfo_FFI on your side?

[quick bough]

I actually inadvertentedly implemented the NKRO interface first and then realised it wasn't working with another keyboard vendor lol

then I had to read the standard

[placid ledge]

HID standard is a big oof yea 😂

[quick bough]

Yeah, I have a macro for the ABI version. This is the struct I use. I can only imagine Rust might have different padding?

[placid ledge]

Could be a padding thing, although it is set to [repr(C)], so it should follow the same padding conventions

It seems to be specifically erroring at the string conversion

[quick bough]

Yeah that Utf8Error doesn't exactly say anything useful lol

[placid ledge]

Specifically this one

[quick bough]

I'm gonna log this value I guess

[placid ledge]

https://doc.rust-lang.org/std/str/struct.Utf8Error.html Seems like it's complaining at the first character it tries to read?

[quick bough]

I don't think the string "Wooting\0" should be that exotic?

[placid ledge]

Would kinda feel like an alignment kinda thing then

[quick bough]

guess I'll need some logging on the Rust side?

[placid ledge]

Would prob be worthwhile to print out the bytes from the pointer it's trying to use on the rust side

[quick bough]

yeah

I did that before I think, but uhhhh, forgor

ah yeah, still have it stashed

[placid ledge]

Must be some real funky bytes it's trying to use for it to fail the utf8 parsing

I don't think I've ever seen that fail before 😂

[quick bough]

well, I think the asserts are fine, still same error

so, probably not an offset/padding issue

Uhh, I set RUST_LOG=trace and now it just crashed outright?

um, does it just call exit(0) with trace?

then it ends up panicking here

Oh, you changed this in your "Cleanup cplugin interface" commit

Well, okay, without that change it seems to be working

but the manufacturer_name is indeed wrong

it still seemingly calls exit(0) sometimes

ah wait I'm an idiot, it's a use-after-free

yeah, fixing that seems to have fixed it

Gonna try again without any changes I made now

yeah, no issues

[strange kite]

Yeah extend shorter presses to 32ms

[quick bough]

I think this approach should be quite future-proof, as long as there's not gonna be further breaking changes 😛

[flat wigeon]

Isn't it a bit bloated to have a second dll duplicating most of the code just to have another version of the abi?

[quick bough]

Well, only one of them will effectively be used, so ¯_(ツ)_/¯

32-bit version of the SDK would be sexy tho, and another reason to ship more than 1 DLL 😛

[flat wigeon]

Who needs a 32 Bit dll? ^^

Do you intend to inject this into old stuff?

[quick bough]

Well, a recent example was GTA SA

It's possible to use analogue input in it with a special build of GInput

but it's all 32-bit, which wasn't an issue for the old analogue SDK written in C

[flat wigeon]

Hmm...

Does the old one work for unmapped keys?

[quick bough]

The old one doesn't work at all anymore haha

[flat wigeon]

Ah right, but did it work for unmapped keys? ^^

[quick bough]

but the firmware interface was based on key positions rather than scancodes from what I can tell

[flat wigeon]

Like in the RGB SDK?

[quick bough]

Yeah

[flat wigeon]

Great

I should have bought it sooner and then never upgrade

[quick bough]

I made a shim for the old SDK: https://github.com/AnalogSense/old-wooting-analog-sdk-shim

I have an OG Wooting Two but I did update the FW

[flat wigeon]

I unmapped the A1-A3 and the Mode key

[quick bough]

You can use F13-F15

[flat wigeon]

Meh

[quick bough]

I did come up with a way the firmware could report position + scancode within 16 bits with some clever squishing, but I guess it's not exactly high prio, lol

[strange kite]

whats the de bounce time on the wooting 80HE?

[quick bough]

I don't think it has any

they do this weird thing when your actuation point is only respected when going down

when going up, it takes a bit more travel until the key is considered unpressed

[strange kite]

I wish you could adjust it

i get double presses when letting go a key

[quick bough]

What's your actuation point?

And are you sure you didn't mess with DKS or RT or shit like that

I once thought my switches were broken because I had RT enabled

[strange kite]

its low but i have it like that for speed running

[quick bough]

Low as in... ?

[strange kite]

now its 1mm so it dosent do that

[quick bough]

Great stuff

[strange kite]

i run a game at 31fps and need a way to set the lowest input time to 32ms

[quick bough]

Yes, you already said this

This isn't exactly the place to give firmware feedback

[strange kite]

i thought you were asking mb i didnt know you already knew

[strange kite]

i put it there

[quick bough]

Might also wanna submit the debounce issue

iirc they were working on something related to debounce but idk

Ah yeah, I found it

@strange kite Go to https://v5.wootility.io/?test-builds=true then you can select a test build in Settings > Updates

Might wanna try v2.10.0-test.debounce.2

[strange kite]

cool i downloaded it

would i just find it as a setting in here?

[quick bough]

As the name suggests, this should fix your debounce issue, nothing more

[strange kite]

oh you cant set it or anything ok

ill test some stuff

[placid ledge]

On that topic of getting key positioning from the Analog SDK, it's something I really want to add (and has been open for a while https://github.com/WootingKb/wooting-analog-sdk/issues/12 😭).

As you can see I haven't had a ton of time to put into the SDKs in general, but it is the first thing I want to solve when implementing a revised analog interface on the keyboard side for the SDK to use,

We're going to be entering somewhat of a transitionary period once we launch the background service, as it will open up some questions about the future of the SDKs and what the best direction is for them. It's something that we don't have all the answers figured out for that yet, but it is something we will have to deal with.

Reason being, the background service will nearly always have the config interface [RGB SDK] and analog interface [Analog SDK] open. The Analog SDK side may be simple if we can get it all working with non-exclusive access, but the RGB SDK will be a tricky one.

For the RGB SDK, it could be a matter of just moving towards having HID LampArray being the primary method, and deprecating this direct RGB SDK access. However, there are still some practical details with HID LampArray that I need to try out before I can say that definitively

[quick bough]

I now have a funny panic when using the example C plugin

[quick bough]

Does this make any sense?

This seems to have fixed it?

[floral crane]

Did something change with the latest rgb sdk update? linux is failing to build on actions now

[quick bough]

Yeah, it's written in C instead of in Rust now

[floral crane]

maybe I have to use ubuntu-24.04 not ubuntu-latest (22.04), it builds on my local linux just fine

[quick bough]

I don't even see the GH actions on your repo

[floral crane]

It's in the WPS repo

https://github.com/ShayBox/Wooting-Profile-Switcher/actions/runs/11537515337

[quick bough]

that seems sus

[floral crane]

you have to rustup default stable after you install rustup

[quick bough]

It's just that shrimple

[floral crane]

what distro?

[quick bough]

See image

[floral crane]

hmm 22.04 works 🤔

[quick bough]

Note the failure before I installed a certain package

[floral crane]

yeah I have that and all the others

[quick bough]

Soup?!

[floral crane]

if you install the exact git version of tauri-cli this needs it uses extra deps for packaging or something, idk

[quick bough]

Damn, gonna have to rename my C++ library now smh

Well, I have no clue about Rust

but I avoid depending on lib___-dev packages by just dynamically linking myself

So, using dlopen and dlsym

[floral crane]

I don't like dynamic linking, things break when you link to different versions, each distro has different versions of libraries, and static linking makes the binary entirely portable

[quick bough]

In this case you're not statically linking tho it seems?

at least I would figure given the failure

if you were static linking, the symbol would be known

[floral crane]

maybe not, I'm not sure how any of this works anymore it's been too long since I got any of this working, it's all hacked together modified old revisions of tauri packages and the old rgb sdk bindings

[quick bough]

Yeah, I'm also not too well-versed with Linux

just recently learned that it's somewhat-common practice for executables to have exports

which can then be imported by SOs loaded in by them

which is like... the exact opposite of how that's supposed to work lmfao

[floral crane]

It made progress switching to 24.04 but now there's an error with a tauri appimage.sh, I'm just not gonna ship a linux build of this update. This project is almost dead unless tauri-egui updates. I can't touch anything or it all breaks beyond repair and I have like 4 repos forked, reverted, and modified just to get it to compile

[quick bough]

Reminds me of my experience trying to get Docusaurus to work just the way I want it

I really don't get all this React crap

[floral crane]

react 😓

[quick bough]

and NPM is such a joke. some packages literally just load another package and redirect calls to them. nothing else.

Yeah, I ended up just recompiling one of the deps and overwrite the built files in node_modules

because apparently source code ≠ build code even tho it's an interpreted language

thanks Microsoft

[floral crane]

I gave up on javascript/typescript it was a mess

[quick bough]

and then a standard MutationObserver to hook into the DOM

because yeah, when you have React with shadow DOM and shit, it's just so unpredictable

[floral crane]

the most impressive stuff is built on a mountain of hotfixes and shit

[quick bough]

nah, not at all

a mountain of hotfixes is making ends meet

literally the definition of tech debt

maybe a bit of sunk-cost fallacy

[floral crane]

I'm talking about javascript

[quick bough]

It's definitely possible to have impressive software that's not a pile of tech debt 😄

but it's very hard in JS because the mindset there is that everything needs to be an abstraction

so obviously everything ends up overabstracted, which is a super easy way to get tech debt, lol

[floral crane]

nothing like writing an entire project in typescript then finding a critical js library that doesn't have types and uses old style syntax like callbacks that you have to figure out with source code

[quick bough]

I don't get the appeal of typescript

it just undermines the integrity of the ecosystem imo

[floral crane]

it just solves a lot of the undefined behaviour at transpile time, in javascript if there's not good docs on everything you might not know something can be null

what does this function take in? look at the source and guess, or look at the typescript types and know for sure, but it's not a perfect solution, types can be separate and not synced with the code, same with doc comments

[quick bough]

Yeah, I would never know without TS 🙏

[floral crane]

but generally it helped me know when I needed to handle something nullable or convert something to something

thats a good function

[quick bough]

[floral crane]

const text_to_text = function(input)

[quick bough]

Guess the return type

[floral crane]

array

[quick bough]

😛

yeah idk, typesafety was definitely not something I thought JS had an issue with

it's more like the semantics of the runtime that are weird

[floral crane]

it is in the less professional libraries, think stuff on github with 0 stars

[quick bough]

Like, '1' + 1 = "11"

[floral crane]

yeah that too

[quick bough]

and last I checked, TS doesn't really help you avoid this

[floral crane]

if you have strict enough eslint/tslint or whatever people use now it does, you can't concat different types without a warning or error

[quick bough]

user_input as number

also imagine if you have a <input type="number" /> and you do .value on it instead of .valueAsNumber, now you have a problem

and I think TS also doesn't really help you here

[quiet root]

the fact its an anonymous function stored in a variable is way worse than anything else

[quiet root]

it does cause it reads .value as string so you do a isNaN(input) and know if it is a number or not

not saying TS is really safe (theres tons of ways to break it)

[sweet peak]

But it's not an arrow function

[quiet root]

sorry anonymous typoed

[sweet peak]

Yeah have fun following that stacktrace

[quick bough]

I do this because I want it to be a const variable (so it's block-scoped as well)

[sweet peak]

you can still give it a name 🙂 its a bit double up, but on stacktraces it will show the actual function name

[sweet peak]

well it didnt used to

good to see that this isnt the case

[quick bough]

I much prefer writing code in Lua/Pluto

which is in fact also why I need to convert between insane (UTF-16) and sane (UTF-8)

[quick bough]

Whooo, monday! Hope you had a nice weekend. 😛

@placid ledge Can't upload a DLL in a Github comment, but this is what I built from my branch on the example C plugin for ABIv1, and it would just panic without that patch.

https://github.com/Sainan/wooting-analog-plugin-examples/tree/abiv1

[placid ledge]

Alright, thanks, will give it a try when I get the chance

[quick bough]

Also opened a PR for this branch now. Afaict, there's nothing fundamentally wrong with this C plugin, and I don't know Rust well enough to understand why this drain-map thingy would turn a 1-sized vector into a 0-sized vector.

[placid ledge]

Thanks, I'll give it a bit of a dig before I do a new release publish

[floral crane]

Tmw the only change was uuid -> uid 😅

[placid ledge]

Tried building this locally on my work machine (macbook) to see if I could encounter the error.
I had to make this change to the plugin.h as there was inconsistency with the function type in the plugin example: https://github.com/WootingKb/wooting-analog-sdk/pull/89

For whatever reason I can't get it to pick-up any device for me, even when it finds one it just fails with no error message lol (even though the logic that is operating now is identical to what the rgb sdk does, which I have no problem using). So I haven't been able to get into the scenario where there is a device in the return for it to be turned into 0-sized vector.
While trying to figure it out, I did make some general improvements to how the loading is done (using usage_page instead of interface number) and some other general bits. I think I'll do those as a new PR once yours is merged.

[placid ledge]

Well. I got it to work. I hadn't noticed that it was using hidapi as a submodule and that version was quite old. Got it updated (& dealt with the bs that came up as a result [Am I the only one that finds cmake a nightmare to deal with?]) and now it behaves completely normally, I don't get any crash with the device info

@quick bough How do you trigger the panic you were getting, just any call to wooting_analog_get_connected_devices_info?

I've got a few commits on top of yours that I pushed to a branch on the repo if ye wanna check them out https://github.com/WootingKb/wooting-analog-plugin-examples/commits/fix/everything/

[quick bough]

Okay, managed to build it, and now it seems to fail even harder

Well, this might be why — it's no longer statically linking in hidapi.

[placid ledge]

In my branch, I changed it to use the cmake build configuration from the hidapi repo itself, rather than the mess it had before

[quick bough]

I can see that, yeah. I am building from your branch.

[placid ledge]

Is it doing dynamic linking for you?

I feel like it should be compiling it fully statically, I think that's what it does for me

[quick bough]

Yes, it's trying to dynamically link, but obviously there's no hidapi.dll anywhere

[placid ledge]

Here I go digging through cmake bs again

I see that it is dynamically linking on my side as well, just that it manged to find the .dll lol

Looks like this needs added set(BUILD_SHARED_LIBS OFF) to the root cmake, will push. It stops it from dynamically linking for me

[quick bough]

Hmm. Still panics, tho.

(I built wooting-analog-sdk from develop)

[placid ledge]

Right, but now it's back at where you were getting it before correct?

[quick bough]

Yeah

[placid ledge]

Just from calling wooting_analog_get_connected_devices_info?

Might be something windows specific

Cus it works completely fine for me on mac

[quick bough]

Yeah, that's the function that's panicking

Like, I have no clue what sorta magic Rust could possibly be doing here (one of the reasons I hate the language), but yeah clearly c_devices mapped from devices (len=1) does not have end up being len=1 itself.

[placid ledge]

The only thing special with the .drain(..) call is that it takes full ownership of the items, i.e. it has more control over when they're dropped. Not really sure how that is coming into play here

With the change you suggested, it would guarantee that the source DeviceInfo won't get dropped until the full scope exits

[quick bough]

So maybe it gets dropped too soon?

[placid ledge]

That's my guess, I'm not entirely sure how that would result in it silently ending up with an empty vector. I've not seen anything like this before. Although, I normally operate in purely safe rust, so undefined behaviour like this doesn't really come up

[quick bough]

Well technically the offending code is not unsafe, nor the variables it operates on

[quick bough]

Like you can see here it's just this map doesn't work

[placid ledge]

True, but there are other parts of the chain that do unsafe things

Not sure how much I want to dig into this. The code change you suggested is pretty valid, so I'm somewhat tempted to just merge it and move on

Although, I do have a feeling that this is indicative of some deeper underlying issue

You only started getting this after the updates to the c plugin?

[quick bough]

Well, I had to update the C plugin to get it to build and run at all

It is weird that I don't have this with my own C++ plugin nor the official one

[placid ledge]

Ye, but you never seen this with your own plugin before?

It is a bit strange that this is only occuring with this example plugin and never happened before

It does kinda imply to me that the root issue is not in this block of code, just that it's the one that ends up failing as a result of it being broken

[quick bough]

I mean maybe it's because the DeviceInfo is allocated directly in the .data section of the example plugin and somehow Rust differentiates there?

Since everything else would have it allocated on the heap

[quick bough]

I didn't see anything else being broken, so idk

[placid ledge]

Well, it is supposed to be converted fully into the rust DeviceInfo type with the wrapper around the cplugin

So on paper, the DeviceInfo's the get_connected_devices_info func works with should all be pretty equivalent regardless of the original source

[quick bough]

Also just tested and it made no difference

[placid ledge]

Part of me feels like the strings have something to do with this

https://github.com/WootingKb/wooting-analog-sdk/commit/6cae8dd47c05e57a1e90544341aad3dd090b0bfc Could you try this out to see if it makes a difference?

[quick bough]

Well, it does seem to have it correctly Rust-ified

Also still panics with your commit

[placid ledge]

Hmmm

[placid ledge]

Getting even more tempted

[quick bough]

I fixed it

Something something invalid states should not be presentable?

Well, pushed that to the C plugin, so techically no issue on the Rust side

[placid ledge]

Oh, huh, that's interesting

I'm surprised that that would result in this behaviour at the get_connected_devices_info level, but nice that it's fixed

[quick bough]

Might be good to have some validation in the analogue SDK to avoid such "invalid states"

but technically the C plugin was at fault for not initialising all fields

[placid ledge]

Hmmm, yeah, it would be good for it to catch this. Right now it assumes that the c plugin has initialised it correctly

Was it just at 0 or garbage data without that line?

[quick bough]

It was 0

although I guess technically it could've been garbage data due to the lack of initialisation

but probably because it's a static variable in the .data section, it was zeroed out

just malloc wouldn't give the same courtesy

[placid ledge]

Right, still confused how this is the symptom we got from that, but anyways

We could change it to just a uint8 for the device_type on the FFI version?

Then we do a more explicit conversion in into_device_info

DeviceType::from_u8(self.device_type) we would then do this call and if it's None i.e. invalid state, we spit out a warning and fallback?

[quick bough]

you would have to make it an int because that's the standard type for enums

but yeah warning and fallback sounds good

[placid ledge]

Ah okay

Although, hmmm, this could interfere with public ABI

With int, it should be the same binary layout

[quick bough]

Not if you use int

[placid ledge]

but maybe it's kinda sucky for it to not reference the enum it's using?

[quick bough]

Does Rust not give you the option to reinterpret an enum as an int?

so, like, device_type as i32 < 0 || device_type as i32 > 3

[placid ledge]

Don't think so, enums in Rust are more than a number alias

[quick bough]

but within the unsafe/FFI context?

so, before the conversion

[placid ledge]

Oh wait, I think I misunderstood slightly what you meant

You just mean getting the number from the enum and validating that?

[quick bough]

Yeah, get the int value in that field

[placid ledge]

Yea, it can be casted into the int value

[quick bough]

On a completely different topic, if your shift ends in 20 minutes, and I'm now going shopping, then this is goodbye for today. 😛

[placid ledge]

Hehe yea, I gotta head soon

https://github.com/WootingKb/wooting-analog-sdk/commit/b61774b2d5f76ce181606e0100d12eed9ab41e0b

Got a little patch for you to try

Enjoy shopping 😁

[placid ledge]

Seemingly it gets sanitised for me somewhere behind the scenes? No matter what invalid input I give it, it ends up with the 'Other' variant. When it's correctly set to Keyboard, I get Keyboard as the output

[quick bough]

I did cargo build --release if that makes a difference

[placid ledge]

Good point, I'll check my build config

Oh yea, I'm defo running on debug builds rn

[quick bough]

This doesn't seem to change the behaviour in any way

[placid ledge]

Welp

Gotta actually head off now, might have some time to look at this again tomorrow

glhf

[quick bough]

hahahahaha, this Rust code actually works if I print the value, but doesn't if I don't print it

Holy fuck, this is AIDS

Well, this seems to fix it (based on your fix/validate-ffi-device-type branch)

If you use an enum type, it seems Rust will just assume it will already be valid and pay no mind to your range checks

except if you print the value then apparently it does?

[placid ledge]

That is pretty wack

After doing some reading, seems like using Rust enum on FFI is just a bad idea in general

So it should just be changed to c_int like you have there

I'll whip up some changes for that in a bit

[placid ledge]

Pushed a change to that branch @quick bough. I did a little trick with cbindgen to get the generated headers to be pretty much the same but it being handled as c_int on the Rust side so we can do the proper bounds check

[quick bough]

Looks good

[quick bough]

Should be close to v0.8.0 now, right? 😛

This one is also funny: https://github.com/WootingKb/wooting-analog-sdk/issues/87

[placid ledge]

Yeah, making builds for it now. Thinking of publishing it as a pre-release so we have a lil window to double check things before pushing it out to everyone on the auto update

Been a long time coming 😅 But at least now I've got the deployments cleaned up, so it should be much easier to push new releases

[quick bough]

I guess I should also publish the ABIv1 version of my plugin then

[placid ledge]

https://github.com/WootingKb/wooting-analog-sdk/releases/tag/v0.8.0

Annoyingly the updater doesn't handle pre-releases any differently, so everyone gets prompted for an update

[quick bough]

Doesn't the auto-updater only run when ran manually?

at least I don't see any prompt when starting analog-midi

[placid ledge]

The updater is only used in Wootility desktop builds

So you get a prompt to update on that

[quick bough]

oh, so no users are affected 😛

[placid ledge]

With this massive changelog 😂

[placid ledge]

You'd be surprised how many people use the desktop builds

[quick bough]

[placid ledge]

Is there any broken UI with Wootility running in Firefox?

[quick bough]

Yes

[placid ledge]

I do know that there are some webapis we used cus they were wider supported than WebHID (but also not supported on Firefox)

[quick bough]

Otherwise everything seems to work fine. Also successfully flashed firmware.

[placid ledge]

Cool

I have a feeling that the UwU light ring display could be broken, but I guess you don't have one?

Analog SDK changelog looking amazing inside of Wootility 😂

[quick bough]

I do have a UwU, yeah, I tested firmware fllashing on it

how could you forget my amazing skills writing at 10 WPM on the UwU: https://www.youtube.com/watch?v=9aKUcVTyI0k

Anyway, seems to look fine to me

[strong siren]

if you do the rainbow preset does the ring look correct?

[quick bough]

Seems to match what I see on the device

[placid ledge]

Ah cool

Works surprisingly well then

[quick bough]

yeah, the gamepad buttons are the only issue I saw

and it's a minor visual glitch I guess lol

[brazen magnet]

The download of wootility does not work from the website

[strong siren]

@placid ledge

[placid ledge]

[quiet root]

replace &channel= with -

just specify the version if you want more than just beta/alpha/stable

the beta.1 in 5.0.0-beta.1 isnt the channel but the prerelease tag which in our case the first part of the prerelease tag (so beta) is the channel

idk if the changelog pages have been updates yet or not

[velvet void]

should be deployed now

[quiet root]

yeah ok now the links look right

[quiet root]

you can retry the link on the changelog page

[quick bough]

woa, more wooting employees I've never seen before

[lament gorge]

they're all on duty on Halloween

[quick bough]

I would hope not.

https://wooting.io/about is just woefully outdated

[lament gorge]

Well at least we can assume they're working on fulfillment instead of updating the website

[quick bough]

Fulfillment is grunt work

I figure most of their staff is for support

gotta deal with people on Discord being like "halp why my keeboardo no workyyyyy?!"

or something like that, I assume

[lament gorge]

lolol very very true

They're still based in Taiwan or Thailand, right? I forget where, but I am sure it'

s just tough dealing with so many timezones

[quick bough]

They're based in Netherlands

[lament gorge]

Oh yeah, right right

idk I thought Asia based on the factory stuff, but I recall their website says Netherlands

not to mention their accents

[quick bough]

Typical Dutch accents lol

but yeah obviously they manufacture where it's cheap

[velvet void]

the week i joined we had 7 new people, i think there's 3 more this monday :p lotta portraits to draw..

[placid ledge]

We're based like ~45% Netherlands, ~45% Taiwan and then ~10% scattered around

I'm part of the ~10%

[quick bough]

Damn, I thought you were Dutch. Is that why you weren't at ClackyCon?

[quick bough]

Well, there are other options, too. 😛

[placid ledge]

Hehe yea, might've been there if I was a dutchy

I'll give you a hint with flags 😛
🇮🇪 🇬🇧 🇳🇴

[quick bough]

You're a nomad moving between 3 countries?

[placid ledge]

haha nah

I live in one and the other two make up my nationality

[quick bough]

I just thought because you're like one of the first employees, surely you must've be Dutch. I mean, you're in the photos, and stuff!!!

https://wooting55730.activehosted.com/cdn-cgi/image/dpr=2/content/mJdQ5/2024/10/31/017db503-e21e-40ba-b40f-4a417df8f034.jpeg

[placid ledge]

That's logical thinking, but I didn't end up in Wooting due to regional locality

but ye, I'm from Northern Ireland but I live in Norway now 😛

[quick bough]

I'm sure there's a story there

[placid ledge]

Well, not super exciting, short version is that I was working on https://github.com/antonpup/Aurora on the side and got in contact with Wooting to add support for the Wooting One/Two to it using the RGB SDK

Around that time I was in University and was looking for an internship, so then I ended up doing one with Wooting where I made the Analog SDK 😛

Then just continued from there

[quick bough]

Damn

If it weren't for you making that SDK, tho

Would've been an even sadder world

[placid ledge]

Just wish I was able to give them a bit more love

But at least I now have more people that care about the SDKs internally 😛

[quick bough]

Are you forgetting I'm not an employee? 😛

[placid ledge]

😩

[quick bough]

I just don't see anyone else comitting to the SDK

[placid ledge]

Not the Analog SDK

but both @quiet root and @strong siren have contributed to RGB SDK

sorry

[quick bough]

hmm

[quiet root]

man i wish i could give the SDKs some better love

[quick bough]

Yeah, I understand that analogue SDK is not a huge priority

[strong siren]

Do I not have an arm in this picture lol

[placid ledge]

Was talking with Erik when he was editing this photo for the newsletter/blog(?) and he was like "no-one will notice that diogo is missing an arm" 😂

[strong siren]

Lmao

[quick bough]

Oh tbh I didn't notice until you pointed it out lol

Camouflage t-shirt

[rocky wharf]

Oh that’s you

[lucid iris]

All u devs are amazing

[quick bough]

[flat wigeon]

I thought that was the main point of the wooting keyboards 🧐

[sweet peak]

Dude is spamming all channels for the giveaway lol

[quick bough]

What giveaway?

Oh, 50 messages

[quick bough]

Well, it's about analogue input, but no games really added native support for analogue keyboards, so we're stuck with gamepad emulation and stuff like SOCD resolution.

[quick bough]

Damn, even I don't qualify

[mental bear]

i have the same problem

lets

start

a

conversation

xd

[quick bough]

Eh, I don't want it anyway, I was just curious if I would qualify

[paper crow]

Do not spam or you will be disqualified

[quick bough]

tbf, you could easily pretend to an active participant by just spouting whatever nonsense comes to mind in #💬│general and sending humorous images in #🐸│memes like normal human beings do with like 0 brainpower exerted

[serene harness]

any update on SDK support for the 80HE? I miss wootverlay

[quick bough]

Yeah, SDK version 0.8.0 dropped, so it's supported now, just need to run the updater

[karmic forum]

Yeah the common misinterpretation for engagement is that it is almost always measured by quantity not quality which has lead to the dead internet theory due to the mass production no zero thought ai content

I will say it is a lot lot easier to measure quantity over quality but still

[serene harness]

can ya build it again for me with 1ms update :3

[quick bough]

Not needed?

Just update your SDK

[serene harness]

Oh

how do i do that

[quick bough]

2 little clicks

[serene harness]

Thx got it :3

i was a bit confused with the git

[deep jolt]

is this the same sdk as on the wootility desktop app?

[quick bough]

Yeah, the Analog SDK is pre-installed with Wootility

[serene harness]

i have a bug on tilde

`

it always shows red (unpressed)

[quick bough]

It might be due to your layout in Windows? Since it's considered to be a dead key.

Might also be my fault because I changed the way the VirtualKey mapping works

[quick bough]

Works on my machine™

and then I switch to a German layout and it doesn't

Well, I'd report this with Woot-verlay, I don't see anything with the SDK being wrong

Actually, it's not even really Woot-verlays fault, since it relies on "NeatInput" to fetch digital input for it?

L

[serene harness]

I'm using an ISO UK board set to UK, and the key works as expected

i used the same layout and language on the Two HE, and it displayed fine

but not on the 80HE

[quick bough]

Well that just sounds weird

[serene harness]

it is not detecting caps lock also

so tilde just appears in red only, and caps lock doesn't show that it's being pressed at all

not a big deal, just reporting bugs

[quick bough]

Yeah, I think you're just gonna have to deal with that

like how I will have to deal with the fact that ViGEm is dead and has no support for emulating touchpad L/R on when emulating a DS4

sucks but I guess I'll just have to take the L and replace my broken microUSB cable instead of trying to get it to work wireless in this stupid ass game that only supports the controller wired

because sometimes that's just how things are

[quick bough]

Okay, so, not exact repro because my 80HE is JIS and my Windows layout is Japanese IME with the underlying layout being US, but also working fine here

[serene harness]

Oh disregard the capslock thing i am just dumb

it's only tilde

[quick bough]

Yeah, I'm like 69% sure it's your Windows layout that's wrong

[serene harness]

and my 80HE hardware is ISO UK

and when i press it, it works

`

[quick bough]

The display language is irrelevant

It's the keyboard layout that's relevant

[serene harness]

that one?

[quick bough]

Yeah, can repro when setting it to that

Complain to the now-unmaintained NeatInput package

Just trust that this random EXE is not a virus 🙏

[serene harness]

[quick bough]

jk, it's a build of https://github.com/AnalogSense/wootverlay-alt-server which I made for Linux

but it also works on Windows, so worth a shot

[quick bough]

130 EUR for a keyboard case 💀

[sweet peak]

This is a dev channel :p

[quick bough]

I'm a dev, I chat here, it all checks out

[quick bough]

Also kinda funny that they only provide this case for the 60HE... I think this is why smartphone companies don't also sell you a case, lol.

[spice viper]

i dont know if ur too stuck up to realize but that case (entire bundle) is made only by request from optimum. you're acting like theyre shoving it in your face telling you that if you dont buy it the world will end.. even optimum himself said he discourages people who wants to buy it for the sole purpose of supporting him (optimum), and that people should absolutely only buy it if they like the product(s).

its an enthusiast product bundle, if you dont want it, dont buy it.

[quick bough]

[gritty lake]

[spice viper]

also about this; yeah no thats not how this works...

[gritty lake]

but 130eu ain't ex

[spice viper]

expensive?

[quick bough]

Seems like someone is happy today

[spice viper]

im just ready to speak when i see someone who doesnt read channels names and descriptions

[gritty lake]

130eu is very decent for custom cases in the market

[spice viper]

my point exactly

@quick bough

[quick bough]

Okay, but this is my channel, I've long-sinced marked my territory here.

[gritty lake]

fr

[quick bough]

Your little pissing here isn't gonna make a mark in the grand scheme of things

[spice viper]

probably the cringiest thing u couldve said to me

you cant just "okay, but.." me and expect me to not laugh at you

[quick bough]

I'm certainly no Wooting hater, otherwise I wouldn't be here, but I think we also shouldn't just eat up everything they do without some fair criticisms

Obviously, the Optimum products are not for me. I'm not gonna buy them.

But the thing is, because it's exclusive to the 60HE, there wouldn't even be a point in me considering buying it, which seems a bit whack.

[spice viper]

hey i dont know if you didnt take the time to, but if you just read what i said then you'd realize that not a single person (who stands for the marketing) is encouraging you to buy the product unless you specifically like that case which is a very high end, nice looking custom case

usually they go for alot more

and if you wanted to provide feedback there are literal channels for that

[quick bough]

I don't want to provide feedback because I don't have feedback

[spice viper]

i commented to see if you truly are as dull as i thought, and i guess you're not, you just like doing shit thats out of the norm

[quiet root]

no

[quick bough]

[quiet root]

you chat about dev stuff here like others. btw. if you wanna continue talking about the case please go to the appropriate channels or #💬│general

[spice viper]

thank you tony :)

[quick bough]

no mini-modding please

I can't believe this guy comes in here with such an attitude and you don't see an issue with that

[quiet root]

i havent really read any of the convo except seeing you said this is your channel which it isnt and you also claiming you chat here cause youre a dev (implying it doesnt matter about what) which is also not true

and reading the convo now it would seem you both have the same attitude problem.

either way this channel is for talk about developing stuff surrounding our products in software by third parties.

[quick bough]

Okay, I'll be sure to not make pissing jokes, lol

On the topic of software by third parties surrounding your products, why is the Wootabase claiming to be "third-party" software but it's hosted at profiles.wooting.io?

I think it's also not on Github, so idk where to complain about there being no JIS option for the 80HE smh

[spice viper]

could be done in #1141096226884956181 :)

[quick bough]

I'm pretty sure that's for first-party things

Gonna post a #1141096226884956181 because my Razer Chroma isn't working

[spice viper]

bro?

"idk where to complain about there being no JIS option for the 80HE"

ill leave this chat i dont dev but man i was just trying to help

[quiet root]

yeah its a sort of odd remnant of being more open to show of community projects. its made by @clear hedge but back then we wanted to show it off more (especially cause we dont have our own version) and thought it might be nice to just give them a subdomain on our main domain. this causes quite the confusion though due to our steady growth and a lot more casual users joining. there is internal talks on how to add this but will prob take some time.

so yeah cop can maybe add the jis option but idk if he will

[quick bough]

tbf I did remember vaguely something about a community-maintained list of profiles for games and then I just never bothered following up on that until now lol

so probably not a huge priority to have if everyone just forgets about it after they buy the keyboard lol

Also discord.io is a dead site now

[clear hedge]

It is a possibility, I'll add it to my list, currently in the stages of alpha testing the next big release... That I've been on and off working on since the last update

[clear hedge]

I take more direct moans 😂 (then I get less)

[flat wigeon]

It is a metal case then, right?

[flat wigeon]

Actually, channels are meant to separate topics, not people

[quick bough]

Sure, because that's totally how that works

like idk how it is for "normal people" but in most of my servers, I keep to one channel, typically the one where I think the smartest people are

even in my own Discord; I have staff to take care of the other channels

[flat wigeon]

On your own Discord server you can manage your channels as you wish. But generally if the channel has a topic you can assume it is meant for this topic.

[quick bough]

That's cool, but there's hardly anyone who rolls like that

[flat wigeon]

That's not true

I can understand why the Optimum Case is that expensive. But I can't understand why you would want it

Also a little bit off topic is not that problematic, just shouldn't be too much

[quick bough]

Woaw there, is this non-community_dev talk in the community_dev channel?

[sweet peak]

-_-

[flat wigeon]

Split channel conversations are bad as well

My issue is not, that you started a piece of off topic here, but the pissing contest thing when you clearly started off topic was a bit much

[rocky wharf]

Guys come on, just keep this channel related to development.

[quick bough]

We're developing a problem here, clearly

[paper crow]

You seem to be developing a problem here.

This channel is not your playground to do as you choose. This channel is to discuss community-led development in the Wooting ecosystem.

If you have a different interpretation of this reality, that's on you, but you are still expected to comply.

[quick bough]

I wrote like 1 off-topic message. Now we have a 50 billion message conversation about how I shouldn't be off-topic chatting.

Partly because everyone seems to like mini-modding

and partly because it seems the actual mods have no idea how real humans actually use Discord

Resulting in a small problem being developed into a bigger problem — so it all checks out

[paper crow]

Perhaps my understanding of your perspective is incorrect, but my point is broadly spoken to defend the organization of our discord

[vital shale]

Hello, I have the Wooting 80HE. I have plugged in the keyboard several times, restarted the PC etc... but it is not displayed. When I am in the hub of Wootility where the demo can also be tested, it says that there is an update for the 80HE but when I initialise again nothing is found. On the 2nd screen you can see that it is found and there is an update. But when I click on it, nothing happens.

[bitter cairn]

same issue here :/

[quiet root]

2.9? you should go to https://v5.wootility.io

the firmware rn is 2.10

[willow carbon]

HEY guys u know why my wooting have delay?

its slower than my last keyboard

and i dont think so it should be

[gritty lake]

Wrong channel, head to https://discord.com/channels/167181566978555904/167181566978555904 or https://discord.com/channels/167181566978555904/1019755933959733258

[quick bough]

No mini-modding please

[gritty lake]

who tf would help them about it here lmao

[quick bough]

No one

but they can figure that out themselves 🙂

[quiet root]

we dont consider people telling others where to actually get help as mini modding.

[quick bough]

I don't think you consider anything as mini-modding, because I never saw you call it out, lol.

[jagged echo]

Calling out people who want to help?
Just let them be...

[quick bough]

Yes, mini-modders need serious help 🙏

[quick bough]

Anyway, going back "on topic," I almost kinda gave up on my analogue input stuff because I just kept finding Chinese crap that seems to mostly be copied from each other, but then I found NuPhy which is also Chinese crap, but at least the firmware is somewhat-interesting and was able to be supported without too much of a problem.

I'm also now at owning probably 20+ keyboards, which is... yeah, definitely reasonable and sane.

[quick bough]

Funny, my WebHID implementation seems to be working better with this device than the one built into Chrome

[flat wigeon]

Chrome didn't strike me as a piece of high quality software

[azure barn]

lol

real

[quick bough]

omg fresh blood

[azure barn]

been here for like 2 days i think

[quick bough]

[cyan saddle]

It would be nice if the analog SDK allowed you to get the user's set actuaion points. I'm working on a game mod that is replacing the movement keys and I would like to be able to fallback to the correct digital behavior for the parts of the game that do not treat it as an analog input.

[fathom garnet]

wouldnt it be better to just set it to something common like 1 or 1.4 for that since on a analog profile the set actuation is likely not what the person uses on digital keys?

[cyan saddle]

In this game movement is overloaded with digital on menus and junk.

I just have it set to 0.5.

[cyan saddle]

CW: gross keyboard

The keybinds in this mod light up the keyboard.
Off: unused
Green: one use
Red: conflicts

[cyan saddle]

And now if you don't have a keyboard it disables itself.

[cyan saddle]

What are the rules for using the SDKs?

[quick bough]

Can't you just get the digital inputs from the OS?

[cyan saddle]

Hm. Maybe? I asked that before I redid the internals of the mod and it should be possible now.

[quick bough]

What do you mean by this?

[cyan saddle]

Like, you need to distribute binaries to use the SDKs. There are rules attached to doing that.

[quick bough]

LOL

As for the analogue SDK, you can just try to LoadLibrary the "wooting_analog_sdk" and then find the symbols you need

You just have to tell users to install it

so no need to worry about possible licence concerns

but I doubt they would sue you for using their SDK

[cyan saddle]

You still need to ship the wrapper and linking against libraries still triggers licences.

[quick bough]

You don't need to ship the wrapper, you can just do this manually

and again, no one would sue you for using their SDK

[rocky wharf]

Licences are just pain sometimes 😄

[quick bough]

That's why you just ignore them

If they try to sue you, you better have good opsec

That's how it works 🙂

[cyan saddle]

Ah, they are both the Mozilla Public License.

[rocky wharf]

MPL 2.0 doesnt look that strickt, i guess if you add the licence to the output of your compiled application it should be fine (for example in a license folder like some other software devs do).

You need to make the source code part that is under MPL 2.0 available to the end user, or at least say where they can find it.

[cyan saddle]

Yeah, this game has a place for FOSS attributions so I'm going to add them in there.

[quick bough]

bruh, you can literally just dynamically link against the SDK, you need to use 0 third-party code for that

[cyan saddle]

Oh.

It goes to a website.

[rocky wharf]

I don't see a part where it forbids you to make your application like MIT while the the library is GPL V2 😄

Yeah or just dynamically link against it 😄

[cyan saddle]

I'm not going to make people download the wrapper and RGB stuff seperatly. The core Wooting API stuff sure since that's what it's designed for.

[quick bough]

Bro, people need to install the Analog SDK (NOT THE WRAPPER) anyway

the wrapper just checks if the analogue SDK is installed and then redirects your calls into it

You can do that yourself, it's like 3 lines of code

as for the RGB SDK, yes, you do have to ship it

but you're also missing the point where a licence just exists to be something you can point at and say "look, we're not doing anything illegal"

[cyan saddle]

I'm just going to use it as intended and respect the license.

[quick bough]

really licences just exist because corpo lawyers like not opening themselves up to being sued

if you use open-source software, you "need" some assurance that the author yields their copyright (or at least lessens it enough for you to use it)

which typically the licence provides, at the cost of requiring you to give some attribution

but realistically, you have to ask yourself if you're actually going to be sued lol

Wooting is definitely not going to sue you for using the SDK

hence you can completely ignore the licences

unless I misunderstand how the world works 🙃

[cyan saddle]

Sure breaking the law only counts if you get caught, but why risk getting caught when it's easy to avoid breaking the law?

[quick bough]

also I doubt you're writing this software on behalf of a company, and if not, I doubt you're going to ask a lawyer, lmfao

[quick bough]

You're not breaking any law

You're operating on the assumption that the SDK wrapper is copyrighted and that the copyright owner objects to your usage of it

That would be the illegal part

[cyan saddle]

If you break the licenses you are.

[quick bough]

the SDK wrapper is such a basic piece of software, you could argue copyright cannot even apply to it

[placid ledge]

The licensing on the SDK's is more to try prevent competitors from just stealing it for themselves. It's not really intended to put significant restrictions on using the SDK's in a third party application

Like I wouldn't want you to be shipping the core Analog SDK stuff, as it's designed to be self managed. Shipping the analog wrapper and the RGB SDK with your application is the intended usage

[fathom garnet]

It also never hurts checking a license when working with someone else’s code, just ignoring any licenses because what could go wrong is really stupid advise lol

[quick bough]

That's why I also advise having good opsec

[fathom garnet]

The advice to have good opsec so in the case of breaking rules you take longer to be found is equally dumb when it takes but a few minutes to check a license

[cyan saddle]

It took me like 5 minutes to put the licences in my project and to document where the sources/binaries can be found with the relevant file paths of said licensed materials.

I was asking on my phone and figured there would be some advice, but hey lesson learned.

[quick bough]

I do typically leave a link to where I found code when I steal it, but mostly so I know who to blame when it breaks

Correctly complying with a licence would require hiring a lawyer which is a waste of money

[cyan saddle]

Or you know, do your best effort and you're even less likely to be sued.

[fathom garnet]

You really don’t need a lawyer when the most common licenses literally spell out copy this file and that’s it

[cyan saddle]

Which best effort covers this.

[quick bough]

Well, copying the text is easy, but it's unclear to me why programs put it in their compiled output

when putting it in the source should be just fine imo

[cyan saddle]

Advising people to break laws and "just don't get caught" is extremely bad.

[quick bough]

and I'm not a lawyer

so nothing I say is legal advice

how about that

[fathom garnet]

That doesn’t make it any less bad because anyone with common sense should know is it’s bad advise

[quick bough]

Yeah I think I don't have common sense

because I just can't stop questioning everything

which leads me to the conclusion that licences are just for vanity

[cyan saddle]

https://tenor.com/view/okay-then-linus-sebastian-linus-tech-tips-thumbs-up-okay-fine-gif-20668596

[quick bough]

Feel free to do your own thinking and come to your own conclusions

[sweet peak]

It's law. It's a legal issue.

[glass umbra]

anyone know if you can change the colors a toggle key changes to? looking to have a color, and then have it change to red when it it is toggled

[clear flint]

go to rgb tab, and click the little white circle in the corner

it'll change it to whatever color you have in the picker

[glass umbra]

https://gyazo.com/669d46a05943e4b0a4697d668f0b7096 that works for f1 and things like numlk but I'm for something like that for a TGL key that is linked to f21 for toggling mute mic

[earnest valve]

does anyone know where the wooting keyboards are made?

like what country?

[quick bough]

Taiwan, I think?

[earnest valve]

okay thanks

[quick bough]

My keyboard has the classic "Made in China" on it

[earnest valve]

so they made in asia?

or like china

[quick bough]

Well, depending on whose dick you wanna suck, it's either made in Taiwan or in China

[earnest valve]

okok

thanks for help

[quick bough]

I say that because China would like you to know that Taiwan is not an independent country

The company itself is based in NL tho

[earnest valve]

what is nl?

[quick bough]

Netherlands

[earnest valve]

oh

alr

thanks

have a good day

[blazing cloud]

You know what would be really cool, a Wooting Wireless mouse with hyperscroll, and their magnetic switches for the buttons as well as software.

[onyx parrot]

..what? 😭

one way to say it ig

[cyan saddle]

Oh wow that's a thing? Crazy.

[nova nexus]

When you zoom in on the wootility website, it doesnt let you confirm in the snappy tappy unless zooming out
nothing too big jst an inconvenience

doesnt let you scroll to it

[hybrid grove]

Is there a way to press the key down and have to keep actuation without lifting it back up and down?

[junior harness]

Is it possible to put the keyboard to sleep via the hidapi? I just want the backlight to turn off when my screen does, though if a key gets pressed I'd like for it to wake normally

the sdk doesn't seem to do anything like this, but maybe the hidapi?

[quick bough]

You would have to figure out how the Wootility changes the brightness on the keyboard and then send that message yourself

but I think just setting it to all black via the RGB SDK should have more or less the same effect

[quick bough]

can you die please

[shadow shoal]

Tf is wrong with you bitch

Piece of shit

[quick bough]

Damn, it's captain obvious

Next up you'll tell us that this channel is not for product support

[shadow shoal]

ur a loser

[quick sapphire]

Nobody's dying on this channel. Now, show me what you have with nuphy's firmware. Got some plans on my own

[quick bough]

All I know about NuPhy's firmware is their rough analogue report format

[quick sapphire]

That's perfect, I can use that

[quick bough]

What are you trying to do?

[quick sapphire]

To make an overlay to show what key is being pressed and how much. Need to show something to them that their DKS implementation is broken

[quick bough]

Already covered:

• https://github.com/DjCrqss/Woot-verlay
• https://github.com/AnalogSense/universal-analog-plugin

[quick sapphire]

Ok that will work. Chances of them getting dks fixed is probably slim to none since they probably just reversed engineered wooting's firmware and don't have a clue on how to fix it.

[quick bough]

Well, their report format was novel to me at least, so it's not the typical copy-pasted trash. 😄

[quick bough]

Ah yeah an alternative is also this https://analogsense.org/JavaScript-SDK/demo tho I personally encountered that Chrome's WebHID implementation was a bit hit or miss with if it does receive NuPhy's reports.

[quick sapphire]

Someone should reverse engineer their firmware so that othera can try to fix it

[quick bough]

Or they could open-source their firmware 😛

[quick sapphire]

Fat chance...and get sued by wooting

[quick bough]

I don't see what for if it's original code

[quick sapphire]

If it's original code. But based on how they can't fix simple stuff like dks releasing a key while physically pressing on the actual key, makes me wonder

[quick bough]

Don't attribute to malice what can easily be attributed to incompetence 😉

[quick sapphire]

I've reported this to them since August, and sent them a reminder every time they release a new firmware, and they always repond with they're working on it

[quick bough]

Well, I'm personally very happy with my Wooting Two HE

Razer seems like the only viable competitor in my eyes, but their hardware and software just don't feel quite right

All Wooting needs is a (new) TKL and they could probably dominate a big chunk of this market

[quick sapphire]

Wooting firmware is solid, found a simple bug and hopefully they'll fix it
#1304193718047342632 message

[quick bough]

Well, aktschually, there is no "key up" and "key down" event in the HID communication

but I personally don't use advanced features like DKS, so I'm not too sure about this stuff

[quick sapphire]

I've been playing a game that takes advantage of nuphy's macro, and got somewhat dependent on it. Having a hard time going back to wooting, perhaps when my optimum order gets in, I'll switch back

[quick bough]

Can't you use Wootomation for macros?

[quick sapphire]

Yes, but I have to map an unused key instead of using a dedicated macro key. Also I can't use wootomation macro to login

[quick bough]

To login?

[quick sapphire]

I lock my computer when I walk away from it

[quick bough]

Ohhh

You might wanna use a Yubikey for that?

Tho I think you can repurpose the A1/A2/A3/Mode keys for macros, just bind them to F13 and up

or using Fn layer maybe

[sweet peak]

Strange use case. You lock pc when not there, but want to unlock with a keyboard macro?

[quick sapphire]

That's my plan if I swith to wooting, use f13-f24 for wootomation macros

[quick sapphire]

Mentioned this to my peers and they also laugh about it

[quick bough]

Probably funniest that it completely defeats the security 😛

[sweet peak]

If you know which key to press

[quick sapphire]

IT wants at least 12 char pw with mixed numbers and special chars, I'll macro that shit

[sweet peak]

Which is quite easy when looking at you once

[quick bough]

yeah fair enough I guess

[sweet peak]

It won't be happy with that. But they do make extreme demand

[quick bough]

or just by him openingly saying "I use a macro key for that" and then try pressing each macro key once

[sweet peak]

And.. Change it every 3 months as well I guess

Anyway the macro isn't stored on the board. I don't know if it'll work through the lockscreen

[quick sapphire]

Sometimes I get a call and they want my password to login, so I tell them to press the macro combination and it will log them in

[quick sapphire]

Nuphy's macros are in the keyboard

[sweet peak]

why we talking about nuphy

[quick sapphire]

Because their dks implementation is broken and they don't know how to fix it

[pulsar cipher]

Good day, a newbie question I guess. On my other profiles the logo for this is lighted on yet here it isn't. Anything to be concerned of? TIA for any answer 🙇‍♂️

Note:
I'm using the beta 5 right now. And it's an active profile.

[gritty lake]

the profile's chosen colour is black

to change it press the 3 dots

[pulsar cipher]

Clicked the 3 dotsa yet nothing happened, it just pop up the options for edit profile. share, etc

[clear flint]

edit profile, then click the pencil icon to change the color

[quick bough]

https://tenor.com/view/ben-shapiro-im-unbanned-gif-22722749

Totally appreciate the irony of everyone treating this like a product support channel but when I get mad that people post off-topic stuff here, I get timed out.

[quick sapphire]

Do Wooting devs, especially the firmware folks, participate in discord? Or are they being tasked with customer service work?

[balmy iron]

They do sometimes

[quick bough]

I don't think they do customer support, but they are here.

[quick sapphire]

The last part was a reference to Cal’s video about staff being tasked with customer service work

Would like to asked them if they’re working on macro in the firmware.

NuPhy has it already and I feel like Wooting is falling behind. #1231685375131390073 message

Give us dedicated macro keys that we can mapped and assign to advanced functions

[quick bough]

Oh, I don't pay that close attention. Funny if true.

I wouldn't want to put developers on customer support, tho. I don't know any good developer who's not either devoid of emotions or just unable to process them normally, lol.

[paper crow]

No no no, I'm going to stop you right there.

You told someone to go die. That is utterly unacceptable. That you don't recognize that is extremely concerning.

Behavior like that will not be tolerated, and if it happens again then the consequences will be more severe than a 24 hour timeout.

[quick bough]

Okay

[quick sapphire]

Damn big hammer was about to come down

[quick bough]

Myself included, of course. At least I recognise that much. 🙂

[velvet void]

which video?

[gritty lake]

calder not calslock

the last one about cs

[velvet void]

oh lol

[quick bough]

I love Counter-Strike 🙏

[quick bough]

derp

[quiet root]

also why is there completely off topic convos again

[quick bough]

idk, why are there tons of people in here asking for product support?

[rocky wharf]

If there would be a way to add it myself, I definitely would try it. For the time being, I created a Flipper App that lets it act as a serial and hid device at the same time (most with native libusb_stm32 stuff). So I can send hid commands or just plain symbols to the device either over serial or hid and then the Flipper send it back to pc over hid. To counter most of the software macro problems. But still does not work for login as it still needs a software.

[quick bough]

Why does it need a software? Is it not detected as a generic HID keyboard?

[rocky wharf]

Yeah it is but in order to send stuff to the flipper zero, I need a software running. It’s just that I don’t send software triggered keys to windows but rather hardware triggered keys basically.

Can’t plug in a keyboard to it. This would need a custom gpio board attached to it and so

[quick bough]

I still feel like a Yubikey is exactly made for this kind of use-case

You can configure a macro on it with the software

[rocky wharf]

What I wanted to do is adjust Wootomation for it to send it to my device instead of sending it as software key.
Yeah but that’s what I have laying around

[quick bough]

and allegedly Windows Hello even supports Yubikeys as a login factor, but I personally never saw an option to enable it

[rocky wharf]

It’s not only for login. It’s to send keys from a hardware device instead of just software.
To compensate the lack of macros on the keyboard itself kind of

[quick bough]

...right

make a device to MITM the HID comms between your keyboard and PC to insert keypresses when it detects a macro sequence being pressed

[onyx parrot]

hello everyone. im wondering how i could change my whole keyboards color with the wooting rgb sdk using Python

[quiet root]

sending a color array

[rich notch]

can wooting utility remap keys for cursor movements (up/down/left/right)? i ask this coz i did it on a qmk keyboard, wondering if i can do the same on my 80he

[sleek maple]

Not directly, but you might be able to map some keys to F13+ and then use AHK to perform the cursor moves when those keys are pressed (unsure if Wootomation supports cursor movement)

[prime rock]

ik this is a bad idea, but surely one of these settings allows me to set rapid trigger lower than 0.15 right?

[rocky wharf]

I wouldn't count on that, there is a possibility that the firmware checks all values to be in the desired range.

[prime rock]

thats too bad

[onyx parrot]

Anti Rapid Trigger
The switch actuates at 4mm and deacruates at 0.1mm.

[quick bough]

Can't you already configure that?

Guess not, damn

[quiet root]

you should be able to by making the actuation point 4mm, setting continuous rt and then make the sensitivity as high as it goes

[quick bough]

This doesn't get any less stressful haha

Anyway, I'm not sure if I'm doing something wrong here, but it seems like the release offset is relative, not absolute, yeah?

[scenic rover]

Is there a way to use the wooting-analog-sdk to get the state of a key which is unmapped in the current profile? If not is it possible to do this using a custom plugin?

I remember asking this roughly a year ago, hoping that maybe something changed

[quick bough]

No, it only reports scancodes, so the keys have to be mapped, but you can use keys like F13 and up

[orchid heart]

anyone know if theres anything to use wooting keeb as an mpc/velocity sensitive pads for music production? theoretically should be able to track velocity of keys so could be used as a pad?

[quick bough]

https://github.com/WootingKb/wooting-analog-midi ?

[fervent mirage]

I need help, I just got my wooting 60he and after a few hours this happening

Only happens with the “1” key

[dusky summit]

Can some one help me with wootverlay

I did everything it says here

[grand flax]

@dusky summit have you restarted? (after installing wootility and opening it)

[quick sapphire]

I think it's open source in github, maybe the developer can chime in

[grand flax]

That's me 😅

[quick sapphire]

🫢

[mortal geyser]

yooo how long does a kb take to get here i ordered it on nov 23 an still havent gotten no emails or msgs abt it

[peak fiber]

Hi everyone, I have a question about customizing the SOCD input behavior on my Wooting keyboard.

Here’s what I’m trying to achieve:

When I press U, the key activates normally.
When I press I while still holding U, the I key should override U (Last Input Priority).
However, when I is fully pressed, I want both U and I to deactivate, essentially stopping any input.
This behavior is the opposite of the "Alternative Fully Pressed Behavior" setting.

Is it possible to configure this using Wootility, or is there another way to achieve this? Thanks in advance for your help! 🙂

[quick sapphire]

Is there a public sdk or api to read and write profile? Thinking about making a system tray app to change and load different profiles

[tawny crater]

strange

uhh i’m like really new too Wooting but i’ll bump your message cause that’s strange.

[quick bough]

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

[sweet peak]

wrong channel, it should be in #1019755933959733258

[quick sapphire]

Thanks for that! Wonder if it can be done without relying on Wootility data

[quick bough]

What Wootility data?

[quick sapphire]

I'm assuming where profiles are store locally. Still reading up on the code and figuring it out

[quick bough]

It only needs the profile index (0-3 inclusive)

[quick sapphire]

It seems to be looking for leveldb

[quick bough]

Oh, hmm

[quiet root]

you dont need any wootility data

[quick sapphire]

Maybe the dev can chime in

[quiet root]

you just instruct the keyboard via the sdk to switch to profile index 0-3

and thats it

[quick sapphire]

Or a contributor 😉

[quick bough]

Well, basically to switch the profile you just need to send a feature report to the device with report index 0 and the data of { 0xD0, 0xDA, 23, profileIndex }, zero-padding the rest

[quiet root]

https://github.com/ShayBox/Wooting-Profile-Switcher/blob/original/src/main.c#L120-L129

[quick bough]

that seems weirdly over-engineered

[quiet root]

it was

[quick sapphire]

It needs to be able to load profile from local storage to index 0-3

[quiet root]

only if you want to load a profile thats not on the board yet

[quick sapphire]

Yes, that's the concept

[quiet root]

and also if you do that youd need to store the profile youre overwriting first otherwise its gone

im not sure how exactly we implemented it internally for our background service

[quick sapphire]

I'm sure that can be worked out. Just need a quick way of loading profiles without opening up a browser or wootility

[quiet root]

read the data from the browser

in a way you wont get flagged by antiviruses

[quick bough]

> write program in C++
> flagged by anti-virus
> mission failed

[quick sapphire]

It'll probably use the same local data the browser uses

[quick bough]

You can probably just record & replay what Wootility sends when you switch a profile in

not that I've ever tried that, but surely it should work