🤖│community_dev

[opaque fox]

I haven't. I could try that. It's not a viable solution though since I will hopefully not be the only one using the thing I'm making. Can't really go ask people to install/compile an untested version of the plugin.

[quick bough]

"untested version of the plugin", so the official one? 😛

I mean, the good news is that I personally haven't had the init take 3 minutes

The bad news is, you're on your own in figuring that shit out

[opaque fox]

Yay

[quick bough]

I made my own C++ plugin, tho it's aimed at supporting keyboard by Razer & co, it can also be compiled with Wooting support, so you could try it as a replacement ¯_(ツ)_/¯

[opaque fox]

Oh? How is it compared to the official one feature-wise?

[quick bough]

It's a plugin, I implemented all the APIs and stuff

https://github.com/calamity-inc/universal-analog-plugin

You'd just have to remove this continue

[opaque fox]

Looks interesting.

[young berry]

hello scary hackers

[quick bough]

Hello, GTA Online DLC name

[young berry]

this name is from the 80s

um, do any of you fine folks know how to get the browser hotkey (in remap) to work when you don't have the browser utility open

[quick bough]

Browser utility?

[young berry]

yeah the beta utility

[quick bough]

You mean this, right?

[young berry]

correct

i have it mapped as such, but it only works when i have a* browser open already

[quick bough]

Well, I just mapped it, closed chrome, and pressing it seemed to re-open chrome

[young berry]

great

[quick bough]

This might be related to how your OS handles this key

[dim flax]

Able to replicate the issue on Arch Linux, works fine while a browser is already open

[young berry]

i was thinking the same thing, do you think this would be a keyboard or windows internal setting though

[quick bough]

It must be some OS setting, the keyboard just reports that the key was pressed

It's possible that the key itself actually has no effect (in your OS) and it's the browser software that implements the functionality

[young berry]

this is why i came to the nerd channel

is there a way to see what the hotkeys in the utility do?

[quick bough]

They're all just keys 😛

[dim flax]

It triggers KEY_HOMEPAGE according to evtest

[young berry]

so it only brings up a homepage for some of us

what os are you on Sainan?

[quick bough]

It's interesting that when I press it while in Chrome it opens up localhost

[young berry]

if you don't mind

[quick bough]

I'm on Windows 10 Pro

[young berry]

yeah same

hm

are we allowed to ping devs?

would this be such an occasion?

[quick bough]

Although, there is also the key that opens explorer and that works fine. That one definitely must be implemented in Windows.

Well, it's not a Wooting issue imo

(Unless you can prove that the keyboard is actually not reporting the key being pressed)

[dim flax]

yeah so KEY_HOMEPAGE doesn't do anything on most OSes afaik, and instead depends on the focused application. If chrome is open and in focus, it'll handle the keypress and go to the homepage, but under any other circumstances it does nothing.

[young berry]

i would just like to know if they would have any insight, i'm not blaming anyone for anything lol

[young berry]

interesting

would changing the function of that key allow all OS to open the default browser?

[dim flax]

yeah

[young berry]

(devs)

right?

[dim flax]

File Explorer doesn't seem to work in Linux either (at least not on KDE), it just sends the KEY_FILE event

[young berry]

feedback channel probably the right place for this?

[quick bough]

It's interesting, tho, I don't seem to get any WM_KEYDOWN or so for it being pressed, it just opens my browser

[young berry]

if i do make a feedback post, would you mind contributing to it? you have a much more in-depth knowledge of what you're talking about than i, lol @dim flax

[dim flax]

might be a difference between Linux/Windows?

[dim flax]

sure

[young berry]

i'm using win10 pr

so /shrug

[dim flax]

It probably just means for some reason that keycode isn't associated with anything on your machine

[young berry]

god these keys are so responsive, so many extra letters!

i guess i want to make sure it's not a ME issue before posting something new

[quick bough]

It might be these launch app keys, since LAUNCH_APP1 seems to be the "open explorer" one

[quick bough]

Well, it is a you issue, since the "open browser" seems to work flawlessly for me

Actually, it's this one

[young berry]

if it's not just me, though

[quick bough]

For the Linux user, you can just blame all problems on Linux

[dim flax]

I can always just rebind it to some macro that does the exact same

[quick bough]

It seems to respect my default browser here

[balmy iron]

Those fancy keys do work on the Linux side, you might just have to bind it in whatever desktop environment / compositor / whatever you use
Check evtest, wev, xev..

[quick bough]

tho I can't find anything else in windows settings about this key

[azure rapids]

hey guys

i am about to purchase my first wooting keyboard

but i have a question to anyone working on the company before purchasing about VAT

does anyone know how i can contact?

[quick bough]

Is this the new general now

Anyway, I'm pretty sure you will pay whatever VAT is appropriate for your current region of residence as per your billing address

[azure rapids]

i'll go to general.

[quick bough]

You'll see at checkout at latest 😄

[azure rapids]

the thing is that the shipping in my country is astronomic, and i can ship to my parents house in another country where the shipping is 10€, go pick it up in 30 mins and go home. but the shipping in my country is like 60€, so the vat and shipping i end up paying in both situations is the same.

[quick bough]

Although it seems like VAT may already be included in the price?

[azure rapids]

look...

[young berry]

thanks guys, appreciate the help. posted a feedback thread and tagged you two in it. feel free to add what you wish!

[azure rapids]

check the shipping and vat difference

[quick bough]

hmm

[azure rapids]

crazy VAT in Spain but ok shipping... the thing is i dont want to pay VAT in spain as i am not living there...

well, do you know how can i contact someone? and sorry to break the theme on this channel, saw it before than general...

❤️

[quick bough]

Well, in my invoice it looks a bit different

Funny stuff

[azure rapids]

it may be different between europe and us...

i have no idea

[quick bough]

possibly

[azure rapids]

thanks!

[quick bough]

they do have a support email, I'd try shooting them a message

but can't you solve this yourself by simply using a different billing & shipping address?

also, Spain is in Europe?

I'm also in Europe

so, uhh

[quick bough]

Holy what, I just deinited the Wooting Analog SDK and it did not crash

lucky me I guess

I feel like I actually prefer a more subtle effect for games, something like (value - 60) / 80 instead of simply value / 255

Guess I'll have to add such options to AnalogSense 😛

[quick bough]

hahahaha, every single time

[safe pier]

40% ortho wen

[quick bough]

nah, they need to compromise more... how about one with only 3 keys?

[kindred furnace]

I currently am using a ZSA Voyager (https://www.zsa.io/voyager). They have a great set of software that enables a lot of neat features on the keyboard.

One of these features is called combos (https://blog.zsa.io/2212-combos/). I really like this feature. For example, if I press the x and c keys at the same time, it returns [ (holding Shift & x + c at the same time, I get { instead). I can similarly press , and . at the same time to get ] or adding shift to get }.

Does anyone know if I can recreate this using the Wootility software for a Wooting keyboard?

[quick bough]

I don't think Wooting have support for this at the firmware level. At best, you can use the Fn key for custom "combos". You'd have to make some custom program that detects your inputs and sends appropriate inputs to correct it.

[kindred furnace]

Ok, thanks for letting me know

[quick bough]

tbf, the software is not that complex, e.g. this here works exactly as you describe:

it's just not particularly elegant since the x & c obviously still get input first

a custom driver could maybe handle this more elegantly, but it would certainly increase complexity 😄

also, a proper implementation of "combos" would certainly add input latency because whenever you press c, there needs to be a short window within which you can press x without a correction being needed (and vice-versa).

[pearl yarrow]

is there a wooting rgb sdk for python or is it not possible

dont really know how it works

[quiet root]

its possible but would just call the dll file via FFI

[pearl yarrow]

do i need to build my own dll using this github rep

[quick bough]

You can just download the official DLL and use it via FFI

[pearl yarrow]

oh nvm its inside

[pearl yarrow]

where can i find the functions of the dll this is the only 1 i know

[quick bough]

https://github.com/WootingKb/wooting-rgb-sdk/blob/main/src/wooting-rgb-sdk.h

[pearl yarrow]

thanx

thanks

aint no way thanks for the help

[quick bough]

[pearl yarrow]

now i can finally make this work much faster

instead of using a program that moves to each button on wootility lol

[pearl yarrow]

cooked something up

osu on the keyboard

[quick bough]

when you said "osu on the keyboard", I thought i'd be like whac-a-mole where a key would light up and you'd have to press it lol

[pearl yarrow]

I just mirror my screen lol

and its osu mania

[quick bough]

I can see yeah

[pearl yarrow]

too little keys for this lol

💀

vertical mode 😎

[quick bough]

that makes more sense lol

[pearl yarrow]

is there a way to make a profile code for wootility

i want to be able to run my code with an image and then paste it into wootility so i can use it without running the program each time

[quick bough]

just open dev tools and create a profile code

then replicate the request that is made

[pearl yarrow]

what is dev tools?
does it track requests made on a computer?

[quick bough]

In your browser. Just press F12.

[pearl yarrow]

oh

i was using the desktop version

will use web one

wow i got it to work crazy

[quick bough]

I just implemented "Rappy Snappy" in software

"Last Input Priority" would be more interesting, I guess I'd have to keep track of which key has most recently been moved down?

[pearl yarrow]

isnt it already a thing

oh ur own implementation

[quick bough]

Well, Wooting users can access it via the beta

but Razer users cannot select which SOCD resolution algorithm to use

and there's other analogue keyboard manufacturers with no such option at all

[vestal jackal]

What do u mean by that @pearl yarrow

[quick bough]

[vestal jackal]

My friend reporter me since I scored more goals than him on football team @vestal jackal

[pearl yarrow]

before i used to get the colors i need for the keyboard and then run a program that moves my mouse to the color text box and then paste the color in and then click the key corrisponding the the color
and i repeat for all
now i can directly set the color to the keyboard using the script
i also made it so its generates a profile code that u can just paste into wootility

[vestal jackal]

So u can pick the colours @pearl yarrow

Can u show me this one time

[pearl yarrow]

i dont have the code for the one that moves the mouse but here is the one that makes profile code

[vestal jackal]

K

So what does it do @pearl yarrow

[pearl yarrow]

takes any image and turns it into a wooting keyboard's rgb

[vestal jackal]

K @pearl yarrow can u make a colourful one

[quick bough]

To whom it may concern, I've just published version 0.2.2 of the universal-analog-plugin: https://github.com/calamity-inc/universal-analog-plugin
I think the list of supported keyboards looks pretty good now.

At least it's a decent ratio of "works with my software" to "waste of money"

(Still pissed at Corsair and SteelSeries for producing the laziest analogue keyboards in existence.)

[quick bough]

AI is such a waste upon this world, can't even listen to this UwU birthday song.

Back in my day, when we wanted a joke song like this, we'd go on fiverr...

[hollow crystal]

thats cool ngl

[rapid pelican]

Is there a way to make a default profile in the wooting software cuz its kinda annoying to change my profile every time i open my pc

[quiet root]

make it the top most profile

this channel is also more for people developing custom projects with our SDKs. for general support its usually best to ask in #1019755933959733258

[serene harness]

@grand flax Hello, i am using wootverlay and love it but it seems to be limited to 60fps. Is there any workaround for higher update rates?

[strong siren]

seems limited to 40 fps actually: https://github.com/DjCrqss/Woot-verlay/blob/799c6cc99e95a7ee790e1eb5ed6cc3d8fffd16d4/Program.cs#L255

[quick bough]

I made a similar thingy to this, but never really bothered making it a standalone piece of software. You can see it here: https://www.youtube.com/watch?v=xIVNbDQNNW4

Can't say about its refresh rate, but it pretty much just waits until the device has an update available and then redraws the window.

Which wouldn't be possible if I used the Wooting Analog SDK

[serene harness]

why not?

40 updates a sec is definitely much too few btw. It looks jerky in realtime (especially without the blending option) and there are issues like going from 0 press to full press in 1 update when really the board was reporting 25x faster than that

[balmy iron]

Analog SDK only has polling, it doesn't give you a way of waiting for updates

[serene harness]

ah that's fine

[quick bough]

Yeah, I was thinking of maybe porting this to using the Analog SDK and making it like a standalone app

I mean, it certainly wouldn't be impossible, I guess I'd just have to memcmp and then redraw

Anyway, shouldn't be too hard to patch the "Woot-verlay" to have a higher refresh rate?

Okay, this bothers me

Also I legit can't detect any slowness

[grand flax]

Oops

If it's too slow u can edit the code and minimise the delay -- I tried to pick a value to not make it have too many requests

[grand flax]

Haha rip I guess my eyes are slow

[quick bough]

mine too I guess

[serene harness]

you can see the issues a lot better with slowmo or frame by frame but they are visible on fullspeed

for example on gameplay it can show a key unpressed, then update once and it's already fully pressed or releasing

it also adds avg 12.5ms, peak 25ms of latency which desyncs it a bit from the video

with a faster report that kind of thing wont happen

[quick bough]

now I'm kinda curious how you'd feel about my tool lol

if you want a random .exe to execute

[serene harness]

you have dm's and friend requests off

[quick bough]

honestly when I just press a key down, it looks instant to me

[quick bough]

I can't possibly tell the transition states

[serene harness]

these are consecutive frames of a 60fps video capture

the key actuates ingame at 0.7mm+ of travel two frames (17-33ms) before wootverlay catches any movement with current delay. These latencies are cumulative, not parallel.

and the key goes from 0.0mm to 4.0mm in one update, when it actually took me probably about 20 milliseconds to press it physically.

on other keystrokes, whole actuations (press then reset) are lost between updates sometimes

[quick bough]

how do you tell the game actuated it?

[serene harness]

new timing indicator on bottom bar and increment on combo count etc

[quick bough]

let me know if it's any better, it should have pretty much no latency as it goes from receiving the HID report to telling Windows to redraw

[serene harness]

will test out in a bit ty (:

[quick bough]

I patched Woot-verlay to be 100hz and I can't tell a difference, guess I'll try 40hz

[serene harness]

I would just yeet it to 1khz and see if there was any problem 😛 falling back like 2ms, 4ms if neccesary.

[quick bough]

I mean I guess it feels a bit slower at 40hz with instant transitions off, but not slow enough for me to say "it's behind", just "it's got a slower refresh rate"

[serene harness]

it is not an enormous amount of latency, just a noticable and unneccesary one. On a tuned gaming system the motion to photon latency is like 10 milliseconds but the latency on wootverlay is up to 35ms

and also there are events that are either lost or poorly captured if you want to do the kind of thing i do and look carefully over your gameplay, how and when you actuated the keys and if you made a mistake with that - the switches can do it, but they need to be polled more often to record truthfully - especially with repeated actuations and rapid trigger.

[quick bough]

ah yeah just recorded it and played it back in slowmo

feels pedantic, tho

(this is 0.1x speed)

the one thing I don't capture is rapid trigger, but I guess I can compile a separate exe for digital input

[quick bough]

may it help you in whatever way

I honestly never used rapid trigger, looks funny tho

It doesn't seem to make a difference with the way I use my keyboard, but I suppose it would help me spam better (if I adjusted my spamming motions)

can't wait to get banned from osu! (I don't play osu!)

[quick bough]

well, it kinda breaks my analogue input stuff because it thinks me applying less pressure to a key means I want to stop pressing it, as opposed to I want the action to happen with less force

[serene harness]

Everybody uses rapid trigger for osu, and for most other games. It's the main transformative feature of wooting boards and analog switches in general IMO

[serene harness]

Yeah, different modes

Being able to do stuff like variable turn rate in a racing game with analog switch is really big as well, but with the state of software at the moment it's limited to a niche

[quick bough]

but didn't they ban the snap tap stuff because it allows for easier spamming? how is this different?

[serene harness]

rapid trigger uses analog data to modify digital input, so it works everywhere that a regular keyboard does - but better

[serene harness]

rappy tappy allows for multiple keystroke events per key press, and for one button to modify the state of another. They're both generally banned

rapid trigger makes the one key reflect your intention with less error, so it's not banned.

If you press a key down and then release it, that is only allowed to send one "press" - and the press has to be sent on key down.

[quick bough]

yeah, I am trying hard to push on this front, but the only thing I can really do is try to improve the Wooting Analog SDK and support non-Wooting analogue keyboards

You will notice the exe I sent will also work with your Razer Huntsman keyboards or Keychron Q1 HE, should you have any of those 😛

[serene harness]

i'm a wooter 😄

what games would you recommend that work well with analog input?

I know that trackmania does

[quick bough]

Yeah, I still prefer my Wooting keyboards, but it's about the point of supporting the keyboards for me

[quick bough]

I'm not much of a gamer tbh. I love GTA V the most, so I patched analogue input into it, and that's what I use mostly.

I put this up on Github as well: https://github.com/Sainan/AnalogSense

[serene harness]

Have you tried the KSP mod? I've heard of it, but haven't used it

[quick bough]

I'm not familiar with the KSP game at all

I'm aware of the mod because of awesome-wooting but that's it

[serene harness]

KSP is really great for vehicle controls. Less excellent for cars because the wheel physics are not amazing, but cars do work and aircraft/rockets are better than anything else

[quick bough]

more native analogue support in games is certainly always appreciated by me! 😄

[serene harness]

https://www.youtube.com/watch?v=vWuOlWK0Z94 this would have been a lot easier for example with analog inputs 😛 there i had to keep some of the control authorities on screen and adjust them with the mouse cursor because it only lets you input like 0%, 20% (with lag) or 100% digitally. The same authorities don't work for different stages of flight. It's also basically mandatory with digital input to tap buttons on and off at a high frequency to average the actual input that you want which causes stuff to jerk around.

[quick bough]

just a shame it's such a niche thing that no games have native support for it

[serene harness]

yeah

[quick bough]

oh yeah I can imagine flying like that with digital input 💀

my analogue patch for GTA V makes planes so much better lol

[quick bough]

and the one game that did dropped it because of how bad the Analog SDK is, ha.

[serene harness]

which?

[quick bough]

Fugl

Like, Wooting listed the game on their website at one point (that link is now defunct)

I'm not sure what exact problem they were facing, but probably fixed with one of my PRs, haha

I'm honestly surprised how people didn't see Wooting when they started out in like 2016 and were like "wait, wtf, my keyboard can sense how hard I'm pressing my keys? this is gonna be huge in games, especially for racing and flying."

because that was basically my exact thoughts lol

Got my Wooting Two and never looked back at digital keyboards

[serene harness]

I saw a video on the original boards, but they had mixed feedback due to some issues and lack of game support

so it went to the back of my mind until everybody started losing their shit about the hall effect switches and rapid trigger 😄

[quick bough]

Yeah, lack of game support was a point of contention, but controller emulation was supposed to be a "workaround"

[serene harness]

then i got to try it, and the hype is real

[serene harness]

Do you play Baldur's Gate 3?

[quick bough]

it's a shame that in the end it's just using the analogue values at firmware level for better digital input, but I guess that's just how it goes

[quick bough]

No

[serene harness]

Games like that move the camera digitally with WASD

it would be nice to have analog

[quick bough]

I'd love to extend my AnalogSense project to every game, but I'm honestly not that good at reverse-engineering lol

[quick bough]

Well, not quite right on the year. I learned about Wooting in this video: https://www.youtube.com/watch?v=1WNI-f6QDPQ

Ah yeah and the RGB SDK still confuses the Wootility

Simon, are you looking for a colleague? I'll rewrite your software in C++ and fix all the longstanding bugs (same thing).

[quick bough]

ah man, the old switches, back when they were still nice and clicky

[serene harness]

That seems fast, but doesn't show the key press depth in digital mode or actuation status

[quick bough]

digital mode tends to do that yeah, it just shows if the key is considered pressed or not

I'd recommend using both exes at the same time so you can see both the analogue state and if the key is considered actuated

I should probably add some option to display both in the same window, but whatever

[serene harness]

https://github.com/DjCrqss/Woot-verlay/blob/799c6cc99e95a7ee790e1eb5ed6cc3d8fffd16d4/Program.cs#L255 could someone compile a few vers of this for me with this line modified to 8ms and 1ms

or tell me how to set up stuff for compiling it if it's easy (i just figure that is probably more effort than someone who already has it set up running it)

[quiet root]

you install visual studio 2022

[quick bough]

8ms

1ms

[serene harness]

ty

[quick bough]

This project should be using .NET 4.5, having to share all this trash is... certified Microtrash 😄

or C++, with no standard library 😛

[serene harness]

It's reading as 0.0% CPU usage even with 1ms wait

and i can't see any problem or reason for me not to just use that, in that case

[quick bough]

Yeah, your CPU can do a lot in a millisecond, lol

[serene harness]

especially this baby

[quick bough]

Cries in instability

[serene harness]

Thanks for the help

@quick bough you can probably get a similar feeling / tell of something being wrong with it like i had if you increase the sleep time, e.g. doubling it. If you're not sensitive enough for these problems to be obvious at 40 updates per second, they certainly will be at 20 or 10. Use instant transition.

[serene harness]

Not sure if OBS still throttles the browser source, but definitely better (:

[quick bough]

You sure are pressing those keys at an accelerated pace

[serene harness]

better vid

and yeah

top players can very very easily do 1.5x my speed, and sometimes 2x with the same level of comfort

[quick bough]

you don't really seem to be taking much advantage of rapid trigger

[serene harness]

you dont see it that well on the display because it's so fast, but generally no when everything is going great rapid trigger isn't kicking in much

when you get faster and less comfortable there is a really common and difficult to avoid mistake where you don't lift your finger far enough off the key to release it before you press again

[quick bough]

my favourite osu! cheat is this: https://www.google.be/search?q=metronome

(I still don't play osu!)

[serene harness]

and rapid trigger basically fixes that issue, to an extent at least (i have it on 0.7mm to mitigate accidental releases and presses)

[quick bough]

I should plug in my Wooting® UwU™ and give it a shot

[serene harness]

has wooting uwu

doesn't play osu

This is what i use with continuous rapid trigger

with other KB's you have to lift 2mm from bottom out every keystroke

with this it's 0.7 mm

which is a pretty big comfort and advantage gain

[quick bough]

Well, from what I understand the most important thing is tapping consistently

aka with the beat

[serene harness]

Yeah but even when you have a good feel for the beat it is difficult to move on the quicker maps

[quick bough]

Stupidly enough, I also have a drawing tablet, but I lost the pen, so it's just a paperweight now

[serene harness]

like 250bpm map with quarter beat streams is 1000 hit objects per minute / 16.67 per second. 8.33 keystrokes per second per finger

i can hit that for brief periods of time like 3-5 actuations, but i can't hit 30 objects in a row at that speed

[quick bough]

So you have to hit an object every quarternote?

[serene harness]

on streams yes

usually people map jumps on half beat and continuous patterns like triplets and streams on quarter beat

and usually maps are around 160-220 bpm

but

[quick bough]

masochism?

[daring shadow]

wait what switches does the wooting 60he+ come with

[serene harness]

lekker v1 60's or lekker v1.5 60's depending on the batch

[daring shadow]

so i would have to buy the gateron magnetic jades seperate?

[serene harness]

if you want jades, yeah

[serene harness]

this is not even considered a hard map any more for like top 100 players https://www.youtube.com/watch?v=5d432cQoT-c

and it's over 16 keystrokes/sec on the streams

[quick bough]

I wasn't even aware this was a thing, I'm kinda sleeping on keyboard modding I think

my keyboard could be sounding so good!!!!

[quick bough]

that's insane. doesn't even look like he's using a wooting keyboard.

[serene harness]

yeah woots are faster

https://www.youtube.com/watch?v=vxTvRcOTH3s

[quick bough]

does this exploit snap tap or what?

or maybe dks?

idk

[serene harness]

No, he's just pressing two keys 11x per second each

[quick bough]

hmm, so just rapid trigger?

[serene harness]

if you dont have rapid trigger (key deactivates after a slight liftoff) then you end up pressing the key back down again before it's reset, and choke hard

[quick bough]

yeah

[serene harness]

when ramping to a speed that you are not as comfortable with

eventually it becomes even a physical limitation because shit can't move that far or swap directions of movement easily in 40 milliseconds

[quick bough]

Wait, osu is open-source?!

isn't that kinda bad for anti-cheat

[serene harness]

this is what happens with dks exploit for tapping on key-up https://www.youtube.com/watch?v=mh24uiU8jOE

A top player showed it on unranked

[serene harness]

I dunno

didn't used to be

[quick bough]

ah it's the next update (codename "lazer")

[serene harness]

lazer has been coming next year since about 2015 (:

did you see that last video?

[quick bough]

wait whaddefack

[serene harness]

that's osu mania

different game mode

[quick bough]

ah yeah I heard of cloutiful

his method abused like DKS iirc, right?

[serene harness]

yes

I think it was using DKS to have key2 output on up instead of down

and tapping both fingers at the same time

so you get one action on key down and one on key up, and have to move your wrist half as much

https://youtu.be/ULFpTUOjpZo?t=40 this is mania :3

i can't play it

at all

that kind of game is kind of a niche but there are gods of it lol

[quick bough]

Welp, probably not horrible for my first time doing this clicky thingy

[quick bough]

I couldn't even process that fast lol

https://www.youtube.com/watch?v=6Oj79WGfQec

[daring shadow]

snap tap is a feature that is going to stay right? would like to confirm just before buying the keyboard

[quick bough]

Just install the beta firmware, enable it, and never upgrade firmware again

[serene harness]

It's actually paradoxically easier to go faster, to an extent. Both in terms of BPM, and having less warning before you have to take actions

[serene harness]

yes

snap tap is questionable from a getting banned pov though. Probably not an issue, but it wouldn't be super shocking if some game banned people for using

[quick bough]

osu! confirmed snap tap is cheating

[serene harness]

for osu yeah

[quick bough]

so just be sure to check with whatever game you're playing if it's considered cheating...

[daring shadow]

i was thinking more towards games like valorant

[quick bough]

probably fine for that

tho the part that confuses me is: doesn't rapid trigger basically already do what snap tap did?

[serene harness]

I don't think snap tap is actually really useful for osu

actually it probably is nvm

but rappy snappy is more OP

because with rappy snappy you can hold key 1 halfway down

and press key 2

and each keystroke of 2 sends a "2" signal when it goes down, and a "1" when it goes up.

so it's doubling your inputs

[quick bough]

well, snap tap is easier tho because you can press 1 key full down and spam the other

[serene harness]

oh yeah

the jiggle strafe thing yeah

[jade crow]

any idea on where wootility will recognise more keyboards manufacturers?

[serene harness]

wootility is the software for wooting kb's

[quick bough]

Wootility is exclusively for Wooting keyboards. What exactly are you trying to do?

[serene harness]

it's not a general configuration software for all keyboards

[quick bough]

also love how sometimes this chat is more active than general and then suddenly new people come in here for support lol

[daring shadow]

also does anyone know if u get a discount for signing up for the newsletter i swear it said u did but i didnt get one :/

[quick bough]

Uhh, there's these if that's of any use?

(also I'm pretty sure I never bought a Wooting 60HE)

[daring shadow]

thanks, but unfortunatley im buying the reg one so those dont work :/ 200$ hope its worth it 🙏 😭

[quick bough]

It is, Wooting keyboards are nice, and very durable

[daring shadow]

👍

rn im on a ducky mini so

[quick bough]

even if you only want to buy it because you think it gives you an advantage instead of an actual interest in the coolness of analogue input, but... no judging.

[orchid falcon]

At this point who is even buying it for the analog

[quick bough]

I mean, I think the analogue is still the selling point

It's just a better way of doing digital input now, lol

which is just a cruel irony

[balmy iron]

Yeah sad, I bought mine for analog keys but that was back in 2018.. Before double movement or snap tap or SOCD was even a thing people thought about

[daring shadow]

just ordered the keyboard it said thanks for the "pre order" are there none in stock rn ?

[quick bough]

Which one did you buy? The 80HE is right now still a preorder. The other ones should be getting out rather quickly, tho.

[daring shadow]

60he+

[quick bough]

hmm, maybe keep an eye on #📦│batch_updates

but it looks like the 60HE+ us already at a stage where they are able to mass-produce it, so it probably shouldn't be too long until you get it

I feel like the Wooting Two HE is the only thing they keep in stock tho lol

Unserious keyboard company, doesn't even have a TKL, smh

[safe pier]

my 60HE+ took 2 days from order to delivery, within Germany

[quick bough]

That is quite fast for Wooting

I bought like 3 Wooting Twos in total and I don't even remember how long they took to be shipped, I guess I'm just remembering how long the preorders take

[quick bough]

That's a win, I guess?

[paper crow]

Please reserve this channel for only community development related conversations, thanks

[jade crypt]

Mine took 4 days in France

And i received it yesterday

Awesome keyboard

[quick bough]

enjoy the analogue input and then despair at the lack of game support

[red wasp]

Can you enable digital input somewhere?

[quick bough]

Your keyboard is digital by default

but the keys are analogue, so you can take advantage of that in various ways

e.g. by defining how far down a key has to be in order to be considered "pressed"

[red wasp]

Right I’m fairly familiar with Hall effect sensors, Ive done a couple projects at work with them, was just wondering that exact point, whether you could set that actuation point. Very cool. Thinking about picking one up then

Thanks for the heads up

[fallen bluff]

my keyboard keep connecting and disconecting when i try to save anythign in the perfomance tab what do i do?

[lusty dagger]

best to make a post over at #1019755933959733258

[quick bough]

Are you using any RGB apps such as Artemis?

[fallen bluff]

nope

all ive got on my pc is pulsar mouse software

and the wooting movement and wootlity

no other rbg apps

dealted razor and steelseries trying to fix it

[floral crane]

Anyone know how fast the leds can update? or how fast the keyboard can update

[quick bough]

Just spam it with updates, it will block/throttle anyway at waiting for the keyboard to acknowledge the interrupt

[floral crane]

would it queue them or skip them?

[quick bough]

The problem is that there is a bug where sometimes the LEDs do not update anymore and you have to unplug and replug the keyboard to fix it. And I'm not entirely sure this bug is related to how much you update the LEDs at all.

[quick bough]

When I say "it" here, I mean whatever API you use to send an HID report.

[floral crane]

I want to make my own program to make the smoothest rgb rainbow animation as possible, but if it reaches a quiet limit and skips updates it would have the opposite effect, skipping colors

I'd be using the wooting rgb sdk

[quick bough]

Yes, that's also an API that sends an HID report for you

So, you will notice it will block, waiting for the keyboard to "ack" the interrupt.

[floral crane]

hmm, I guess I need to make sure not to overload the processor so it doesn't fail to send keys or something

[quick bough]

I don't think that's a concern

My only concern would be about that bug where it stops updating the LEDs

Anyway, just try it and see 🙂

[floral crane]

for some reason I cant increase my brightness above the locked limit unless I use artemis, this might be a beta firmware issue though

[quick bough]

Wdym?

[floral crane]

it only goes full brightness when artemis is running, otherwise its stuck at 70% even if I set it to 0

[quick bough]

70% is the brightness you set in wootility?

[floral crane]

I tried both wootilitys but its the same, for some reason it only applies when using third party rgb, with stock rgb its stuck at 70%

70% is the max unless you click to unlock the 100%

[quick bough]

Oh, because of Tachyon mode?

[floral crane]

it saves the brightness setting, and uses it when I open artemis, and shows in the other wootility, but it goes back to 70 when I close artemis even if it's set to 0 it goes back up to 70

[quick bough]

I don't think an app using the RGB SDK should touch the brightness at all, as you can configure this yourself in the Wootility

and I figure the brightness cap is only implemented in the wootility because they found it to be better for tachyon mode performance

if you then go and bypass that by sending a report to the keyboard to update the brightness yourself, that's kinda your problem

[floral crane]

I'm gonna try factory resetting the keyboard to see if it's a beta firmware issue

now its working, must have been a weird bug

[floral crane]

I'm getting hidapi error: Broken pipe

I'm using the hidapi rust bindings, and I tried setting my device to 777 with udev rules but it didn't help

[quick bough]

You did set the udev rules appropriately, yeah?

Otherwise the fopen or whatever call returns an invalid handle value (-1 I believe)

Actually maybe any negative number would be an invalid handle on Linux

[floral crane]

I have the wooting udev rules and I tried adding a 777 permission to it

[quick bough]

You also restarted the service after setting the rules?

The chmod has no relationship to it, I don't think.

Tho it should definitely work as root

[floral crane]

no I set MODE="0777" for all the rules, restarted, and I tried running the binary with sudo

[strong siren]

What is this "smoothest rgb rainbow"? Just out of curiosity

[quick bough]

That still makes no sense to me, because this is all the rules you need

[floral crane]

yeah thats what I have

[quick bough]

Well, okay. Have you tried using hidapi in C/C++?

[floral crane]

I want a smooth rainbow animation but with tachyon, but artemis lags sometimes and I want to program some integration with my window manager

[quick bough]

It's unlikely the Rust bindings are wrong, but I can't reason about your Rust code, there's like so many uncertainties about it that I assume you just hope magically work.

[floral crane]

I haven't, but my wooting-rgb-sys crate works

[strong siren]

You lose the benefits of tachyon if you use the rgb sdk :p

[quick bough]

Why don't you just use that one as the framework, then?

[floral crane]

:o

I've been missing tachyon this whole time

[strong siren]

Yeah

I saw people say that too

[floral crane]

the rgb sdk doesn't expose the rgb matrix buffer, so I wanted to try using hidapi

[quick bough]

There is no way you can possibly notice a 1-2ms difference in input latency

[strong siren]

So I asked Simon and yeah, turns out reading the rgb stuff on the firmware side and writing to the leds takes time

[floral crane]

no I don't care that much about tachyon

does it turn back on if you close the program?

[strong siren]

It doesn't turn off

Just the stuff that it skips doing while it's on, you're forcing some of it to still happen if you use the sdk

That's my understanding of it.

[floral crane]

So there's 8 interfaces and #2 seems to accept write/feature reports, but I don't see where that's controlled in the rgb sdk

[quick bough]

You need to check for the usage page

I think it's 0x1337 for the RGB interface

[limber glade]

Based

[quick bough]

I hate to say it but I found another bug in the Analog SDK

[flat wigeon]

Would be nice if you could read unbound keys

[floral crane]

What's the difference between usage and usage page?

[quick bough]

It is more so for organisation, same as why vendor id & product id are split.

Of course, ultimately, it is up to the vendor to decide how they assign usage ids and usage pages for custom protocols, as long as they avoid any of the reserved/documented values.

[limber glade]

Is the firmware open source?

[quick bough]

No

They gotta keep (what they believe to be) their competitive advantage

In my opinion as a customer of analogue keyboards, the real competitive advantage of Wooting is the long-term support they give for their boards plus the excellent software interface (Wootility).

Because as we can see, when a competitor actually cares (Razer), they are perfectly capable of replicating everything on the firmware end that's interesting

[limber glade]

For sure, wish they were hiring US developers. Seems like a great team

[quick bough]

Meanwhile I'm just trying my best trying to fix up their Rust software as a C++ developer

even tho "at work", I've now stopped using the Wooting Analog SDK in favour of my own C++ class for analogue keyboard support. It's a lot easier for the customer to use (no need to install any software, no need to restart anything for PATH to update, no need to install a plugin for non-Wooting keyboards). Also the turnaround time on getting bugs fixed is faster by a factor of months or even years.

[limber glade]

I haven't looked at the SDK yet, might have to take a peak

[floral crane]

I have like 10% of the rgb sdk rewritten in rust, took awhile to figure out the rgb matrix is converted le

I can't spam send as fast as possible, it causes flickering, but 15ms is good enough

[quick bough]

You are waiting for a response from the keeboard, right?

[floral crane]

oh yeah I guess I can skip reading the response

its empty anyway

[quick bough]

You should not do that 😄

Wait for it to acknowledge

I've not managed to have it "flicker"

[floral crane]

do you mean wait for the write to finish? or read, read just timed out after 1s for me

[quick bough]

Well, depends on what you're doing

If you're doing the full-size keyboard update (WootDevRawReport), then I don't think there's a response

but (WootDevSingleColor) does have a response

Init & Deinit as well

[floral crane]

https://github.com/ShayBox/Wooting-Integrations/tree/master/src
It works!

[quick bough]

Now rewrite it in C++

Also, note that your product matching will breaky if you switch the gamepad mode

[floral crane]

yeah a lot of it is hard coded, I might add more support to the lib, but it's really just for me. I will never write in c/c++ again .-.

[floral crane]

does anyone know what command is used to get what function layer is active?

[quiet root]

we have a command for that?

[floral crane]

wootility is able to see it, I assume there's a command for it

[quick bough]

You mean which profile?

[floral crane]

which function layer is active

[strong siren]

where can the wootility see that?

you can get which profile you're on, but i don't think you can for the current fn layer

[leaden vessel]

yay a Rust enjoyer

[quick bough]

If you like Rust, you might also enjoy: The Wooting Analog SDK

(Just ignore all the bugs that the version written in C never had, and then you can feel a sense of pride and accomplishment!)

[leaden vessel]

It was written in C before?

[quick bough]

Yeah, there was a version written in C

Then they rewrote it in Rust, like you do

I don't quite get what possessed them to skip over C++

[leaden vessel]

The fact that it's not safe :p

[quick bough]

That's a strawman argument

[leaden vessel]

Well, that could be one reason

There could also be Rust's package manager or its stdlib and ecosystem in general

[leaden vessel]

what about that is strawman 💀

[quick bough]

It's pretty hard to have memory safety issues in C++

[leaden vessel]

Have you seen https://security.googleblog.com/2022/12/memory-safe-languages-in-android-13.html

[quick bough]

[leaden vessel]

I really don't think it's hard to have memory safety issues with C++. You can still blow your whole leg off while carefully using all the standard practices like unique/shared ptr, all the sanitizers like ASAN/BSAN/LSAN etc.

Meanwhile in Rust:

• when you don't use unsafe, it's impossible unless you exploit a compiler bug
• when you do, the unsafety is scoped and way more controllable

[quick bough]

Different requirements, I suppose

The problems I have with the Analog SDK boil down to the fact that Rust loves crashing

Sure, you're not having a memory safety problem if your program isn't running

but you're also not having much of a user experience to speak of

[leaden vessel]

That's not inherent to Rust, but more about the general practice of using .unwrap() everywhere when starting a project. At some point when I write a program in Rust I want to make sure I have the least amount of panics possible, at which point I look for these .unwrap() / panic! calls and start properly handling the errors

[quick bough]

Also, a lot of these crashes I've seen were, in fact, not panics, but access violations

[leaden vessel]

Wait, they were?

Hold on lemme look at the issues

[quick bough]

I mean, there's the unstable ABI

which funny enough does not cause linker errors

so it links just fine, but crashes at runtime

[leaden vessel]

Are you talking about the analog SDK?

[quick bough]

Yes

[leaden vessel]

Because I searched "segfault", "segmentation", "fault" and "violation" and I see no issues

Did someone get one on the Discord or something

[quick bough]

There was one where a dependency of the package was super unstable (I mean, okay, they used unsafe, but it was not clear at all what caused it)

Like "yeah your code is safe unless you use unsafe" kinda breaks apart when just any one of your 100 deps has to use unsafe lol

With this npm-like ecosystem, how could you possibly tell

[quiet root]

most of system resource access is unsafe

[leaden vessel]

Ah, there I see one potentially related issue, which is fair yeah, Rust has an unstable ABI (I don't like that either)
https://github.com/WootingKb/wooting-analog-sdk/issues/71

[quick bough]

The fix for what I just described was just merged recently: https://github.com/WootingKb/wooting-analog-sdk/pull/70

Still a problem in the current latest release tho

Another fun random crash: https://github.com/WootingKb/wooting-analog-sdk/pull/76

The patch for that doesn't even make any sense, lol

[quiet root]

if only rgb sdk was rust already with proper multi device handling and not my BS implementation of it

[quick bough]

Yeah let's write more code in Rust so we can have more unstable software!!!

[leaden vessel]

Not really. The standard library uses unsafe all over the place, but they actually put comments every time to justify why what they're doing is actually safe.
One of the problems with Rust is that it's harder to write unsafe code in it than to write inherently unsafe code in C/C++ (or Zig, I've heard) because of the differences in UB

Even though Rust makes tangible improvements to memory safety and actually makes it possible to write completely memory-safe code, if at the end of the day you don't look at what the dependencies do you're bound to have any kind of error including memory unsafety ones

[quiet root]

i dont think rust is inherently more unstable or something. also cant forget that c/cpp had a few more days in the slow cooker compared to rust

for how young rust is its pretty mature imo

[leaden vessel]

Writing stuff in Rust vs in C won't make the software crash more than it would've done. Usually it makes it more correct simply because its APIs are designed to make you think about it earlier on. That's my experience

[quick bough]

My experience is with C++ (not C, C should die already) and RAII makes it stupidly easy to deal with allocations, without leaks or GC (same thing)

[strong siren]

"c should die" take is kind of insane lol

[quick bough]

Certainly I will agree there is some "skill" in identifying when C++ code is just C in disguise and then double- and triple-checking the correctness of all operations

[quiet root]

yes but cpp is like 40years old now having the same time in development of the language, its compilers and resources around it

rust is not even 10years old

[leaden vessel]

the first stable release of Rust is 9 years old indeed

[quick bough]

I mean, not entirely, obviously it has tons of uses for tiny compilers/embedded systems, but systems programming on modern operating systems? No, don't do that.

[leaden vessel]

Also people put Rust up to really really high standards when it comes to memory safety because it's pretty much the only language that brings "memory safety" as a point. Garbage collected languages already brought in safety at the cost of performance, the real difference is that Rust eliminates that cost (and also has a few more correctness features)

[quick bough]

even worse when people act like it's either "write it in C" or "rewrite it in Rust" when the correct answer is "slowly introduce C++ to the C codebase"

[leaden vessel]

Have you heard of https://github.com/EpicGamesExt/raddebugger 👀

oh and also Handmade Hero

[quick bough]

No, looks like a horrible language choice for a GUI program tho 😄

[quiet root]

weird cause i remember some kernel devs saying cpp sucks for kernel development

[quick bough]

I know why Torvalds doesn't like C++, and I understand those reasons, but at least some of them should also prevent him from allowing Rust 😄

[leaden vessel]

It looks like it, but the author Ryan Fleury has a whole UI blog series on making a UI from scratch, as well as a few other programming-related articles you might like to read as a C++ dev :p

As well as his Enter The Arena conference talk

Personally I prefer C to C++ because it's a simpler language, which is weird coming from a Rust guy but it is what it is

Ideally I'd have a language that's safe like Rust yet kinda simple like C, but that language just doesn't exist (no, Zig is not safe like Rust)

[quick bough]

Ideally I'd have C++ and more humans with enough brain capacity to effectively use it

and, as a result, more software that is up to my standards

but then, I guess there would be very few opportunities for me to compete and earn money 😄

[leaden vessel]

"skill issue" is a very common argument, but I think it's a complete red herring

[quick bough]

Well, either you never tried C++ or you have a skill issue

and surely you wouldn't dislike C++ if you never tried it? 😄

[leaden vessel]

I did C++ before doing Rust

I even did Cuda

[quiet root]

i hate c++ after trying it

lol

[quick bough]

Well, it's what we use to make complex software without going insane

[leaden vessel]

You're gonna encounter skill issues in any programming language, especially low level ones.
In C you'll get skill issues doing memory management yourself in a safe way.
In C++ you'll get skill issues doing RAII and ownership in a safe way with smart pointers.
In Rust you'll get skill issues fighting with the borrow checker.

[quiet root]

in JS everything is a skill issue

[leaden vessel]

The key difference in these skill issues is that in the first two, you get skill issues making your program safe, while in Rust's case, you get skill issues compiling your program because the compiler only accepts safe code

[quick bough]

(I'll be going afk for an extended period of time now.)

[leaden vessel]

(note that I didn't say it only rejects unsafe code, that's different and somewhat impossible unfortunately)

[leaden vessel]

I love how my first proper discussion on this server is a Rust vs C/C++ debate lmao

[strong siren]

there's been a few of those already lol

[leaden vessel]

I guess this checks out

[leaden vessel]

I'm sure xD

[strong siren]

rust does have a lot of that doesn't it

[leaden vessel]

Yeah

At least I like it on Discord way more than on Twitter. On Twitter they just babble without listening

"Rust doesn't prevent all possible memory leaks on earth?! yet it claims it's memory safe?!?!? ABSURD!!!!!!"

[strong siren]

the only significant thing i did with rust used a lot of win32 stuff that i couldnt find any rust bindings for so it was kind of miserable with a bunch of unsafe blocks everywhere

sad

[leaden vessel]

Yeah FFI is always unsafe u_u

[balmy iron]

I tried Rust years ago and didn't like it and don't like it now either

[leaden vessel]

I start liking Rust less and less for quite different reasons

The first is that it's now really complex and that I wish I could use a safe language that's closer to C in terms of simplicity

[strong siren]

i think rust can be good but it's also misused. people try to shove a systems programming language where it doesn't belong because of hype

[leaden vessel]

The second is that I got... uh... "arena-pilled" (?)

[balmy iron]

I also find it quite offputting that many Rust written projects make it a big deal that OOOH WRITTEN IN RUST YEAH!

[leaden vessel]

The third is that everyone uses a gazillion dependencies and it frustrates me to no end

[leaden vessel]

I don't, I do it myself. It's pretty much a feature in and of itself because of the work the language does. You can expect both performance and memory safety when you see that

[balmy iron]

That's combined with the fact that a lot of those projects seem to crash (well, panic) all the time

[leaden vessel]

Yeah at the beginning they'll be lose on that

[strong siren]

there's shit projects written in rust and amazing projects written in "crappy" languages. the tech doesn't matter that much

as long as it's appropriate for the task, of course

[leaden vessel]

Well, it does matter, but it doesn't do all the job

It's just a tool like any other with its advantages and disadvantages

Not really making a "everything has its ups and downs" argument because I will rank a tool against another one

But talking in binary is always easier than explaining every nuance

I personally try to use .unwrap() the least amount possible in my projects anyways, I prefer when stuff doesn't crash

[strong siren]

that's how the "blazingly fast" people come across to me, binary :p

[leaden vessel]

How about quantum bits? Being memory safe and memory unsafe at the same time

cve-rs wink wink

[balmy iron]

I'm also not too impressed by all the dependencies. I get it yeah you don't want to always write everything but seeing even simple projects have hundreds of them, many even duplicated in different versions isn't so great

[leaden vessel]

lol I see it got shared here too
#🔌│tech message

awesome

[strong siren]

i was looking for that earlier but couldnt find it haha

[leaden vessel]

I actually prefer that Cargo duplicates dependencies, rather than refusing to compile or bring a dependency because it exists in another version

I don't like that the ecosystem is essentially drugged on deps though I'll give you that

I try my best to avoid it but it's hard

[strong siren]

the worst part of c/c++ in my personal experience is build system and dependency things. i get that it's skill issue and the flexibility can be a good thing, but i like to be able to cargo build and it "just works" most of the time

[quiet root]

but what about my dependency on the package to tell me if something is odd or even and the one for if its true

[quick bough]

How could you possibly f*** up RAII and ownership?

Worst case scenario, you use-after-move a smart-pointer and it turns into a nullptr deref

[leaden vessel]

...you just gave an example, didn't you

[quick bough]

Which you can imagine as a "panic" if you wanna think of it as all sunshine and rainbows

In both cases, you don't have a memory safety problem

You still obviously have a logic error, tho

And no amount of compiler magic will fix that for you

You need to know programming to write correct logic

[leaden vessel]

winit, a windowing library, imports about 100 dependencies including ones like ttf-parser, tiny-skia and ab_glyph just for wayland because Gnome is too stupid to enable server side decorations

Transitively, through sctk

u g h

[leaden vessel]

duh xd

Even Rust doesn't change that of course

it just makes it easier to not fuck up

[floral crane]

the real reason I use Rust is to not have to deal with learning gcc/clang/make/cmake and writing makefiles that works cross platform. The language is alright, I would learn it if it was otherwise easy to use; but it's not unless you're writing software for one platform.

[quick bough]

Fair

I wrote my own build system to use C++ cross-platform lol

[leaden vessel]

build.sh + build.bat my beloved (?)

[quick bough]

but the main money maker is Microsoft Windows, and there Visual Studio is an excellent build system

[floral crane]

excellent you say

[balmy iron]

Yeah it really is excellent when I have to manually edit all the library and include paths since no one invented pkgconfig on Windows

[quick bough]

I could've saved myself a lot of pain if I didn't want to port my software to Linux and MacOS probably

[floral crane]

seriously windows where's the pkgconf

[quick bough]

I mean, it has tons of problems, but on the point of pressing Ctrl+B and compiling what needs to be compiled and linking what needs to be linked, it is very good

[quick bough]

Like, secure sockets are just an OS library on Windows

instead I spent probably 2+ weeks to implement TLS myself

[floral crane]

why

[quick bough]

so it works on Linux and MacOS, too 🙂

[floral crane]

openssl 🥲

[quick bough]

very subpar software

tons of memory leaks

[floral crane]

can't argue with that, I use rustls

[quick bough]

(which is not spec compliant)

[floral crane]

and yours was spec compliant and had no bugs?

[quiet root]

obviously

[quick bough]

It was spec compliant, and had none of the known implementation bugs

[quiet root]

its sainan youre talking to

[quick bough]

but it's a protocol with 2 sides and I encountered some funny behaviours from the other side

[leaden vessel]

Also, use-after-move wouldn't be a nullptr deref, it would just be an invalid deref, right?

[quick bough]

e.g. NGINX would return internal_error with certain "hello" packets

[floral crane]

I have been stopped by tls cert pinning too many times

[quick bough]

No, it would be guaranteed nullptr if the value was moved away. You have such guarantees with C++ classes.

[leaden vessel]

Ah right, because it nulls the pointer

[quick bough]

I don't actually use cert pinning, I just have my own trust store, which is fine, unless one of the trusted CAs goes rogue.

Also means you can't MITM the traffic, because none of your puny CAs are trusted

[floral crane]

thats my problem

[quick bough]

Not a problem for my use-case, lol

[floral crane]

I wanna sniff that traffic

[quick bough]

and I don't want you to sniff my traffic 😛

[floral crane]

does that mean stuff that uses cert pinning doesn't work in china, or do they allow tls traffic through the firewall?

[quick bough]

Mainly, the TLS spec does advise that AES-CBC-128 should be supported. In practice, of course, this is rarely the case anymore. This does lead to some problems where people just try to optimise for some security vanity metrics with their TLS configs and end up making it needlessly complex to just establish a connection.

Sometimes people forget that public UDP services exist, too, I feel like

[quick bough]

Not really. Who you connect to is obvious from DNS queries which may be unencrypted, from the IP address you are talking to, and/or the TLS handshake which also includes the target hostname in plain-text.

Funny enough, they block raw.githubusercontent.com, but they don't block jsdelivr (who also support github content)

Trying to make internet-connected software is actually a nightmare with everyone and their mother trying to play the censorship game

I include anti-virus/endpoint security/router security/etc. as "censorship", too, because your intentions rarely matter

[floral crane]

https://tenor.com/view/cat-gamer-cat-cat-computer-computer-cat-kittens-gif-19682674

[leaden vessel]

don't mind me, just adding salt to the Rust vs C++ debate (?)
https://lobste.rs/s/0akqy7/are_there_any_good_reasons_use_c_over_rust#c_ihruks

[quick bough]

Yeah, there's a lot of "interesting" takes on it

I'm not surprised that this is a decision that came from the CTO (and not someone who actually codes all day)

Rust totally makes sense as a business decision

If the mistakes of your programmer can not be a safety risk for your company, then obviously you would prefer that language

because now you can hire more idiots

have a higher throughput

and (at least in theory) not take on any more risks

but the way this scales is also interesting to consider

because a person using Rust is certainly slower than a person using C++

the C++ dev doesn't have to contend with a ton of (mental) overhead everytime they wanna add a new feature or whatever

I'd always just invest in better people over more people

[leaden vessel]

This is yet again the "skill issue" red herring which only leads to unjustified gate keeping. No matter if you're the number 1 driver, you wear a seatbelt, and if there's one that's better engineered than the last you will wear it given the opportunity. Switching to Rust doesn't necessarily lead to having more programmers that are overall less skilled, it just leads to more safety compared to C++

[quick bough]

Your analogy is a bit off

Also, as someone who doesn't have a driver's licence, the only times I was behind the wheel and driving, I in fact did not wear a seatbelt 😛

Anyway, if there were a better seatbelt available, doesn't mean I would install it if it were effort and the benefits aren't obvious

[leaden vessel]

They do. They think about ownership and lifetimes without the help of the compiler to verify it. It has to stay in their heads, just like someone coding in a dynamically typed language has to keep the types of the variables in their head in mind to avoid having runtime type errors

[quick bough]

oh, you might be doing C++ wrong then

my man, that's what abstractions are for

I never think about lifetimes nowadays, I just trust in the architecture I made like 4 years ago

which is a very simple tree structure

the lifetime guarantees are obvious

it's maybe 1 second of thought

but if I had to explain what I am doing to a compiler? that how if X is true then also Y is true (and these might be entirely unrelated from its POV), I'd probably go insane

[leaden vessel]

A simple tree structure is exactly what the ownership model makes you create. When you need something more complicated that's where you use different memory containers (equivalent of smart pointers in Rust, somewhat). They're also abstractions just like C++ has abstractions, they're just better

[quick bough]

So, it is a skill issue

The skill of programming apparently suddenly disappears from you when faced with C++, or what?

Nothing stopping you from making a tree of smart pointers in C++

you have RAII make sure you don't leak

and very clear guarantees about the lifetime between nodes

[leaden vessel]

I'll just quote this, I don't know what you're saying anymore

[leaden vessel]

RAII only prevents dangling pointer memory leaks, not other kinds of leaks, just like Rust with ownership

[quick bough]

If you can show me an example of where you go wrong with RAII and smart pointers, it might make sense

[quick bough]

What kind of leaks are you worried about?

[leaden vessel]

reference-counting cycle is a classic example

another very very very simple one is just pushing to a vec in a loop

[quick bough]

Does Rust solve that?

[leaden vessel]

No, because no language does

It would involve solving the halting problem

[quick bough]

So, you need to know programming either way 😄

[leaden vessel]

which is why it's a waste of time to try and fix that

[leaden vessel]

yes

Did I ever say you didn't?

[leaden vessel]

I saw this video a while ago. I think that's a good example
https://youtu.be/qeiRGbYCD-0

[quick bough]

Okay, at first I thought this guy was being serious

Then I spotted he's returning a reference to a local lmfao

Yeah, it helps to know how to code in C++ for sure 😄

On another note, vector<Node> doesn't make a huge amount of sense

realistically, Node is an abstract base class with tons of implementation details being unspecified

so you do generally have a vector of smart pointers

[leaden vessel]

Of course it helps

[quick bough]

of course, you have to be careful where the lifetime of your objects ends

[leaden vessel]

Yeah, because the compiler won't do it for you

[quick bough]

you can't just give a reference to some value you put on the stack during the function execution

[leaden vessel]

You can, that's what will make it crash - although I'm sure for cases as simple as the first one in this cideo the tools can catch it

It's not always that obvious, but unfortunately I already ran out of examples. I have another one but it's your previous "worst case scenario" lol

[quick bough]

Well, you can, but you can't. It doesn't make sense.

Also in general not an issue if you ask the compiler to infer the type for you

which he already kinda did but then didn't just to invoke UB for the sake of invoking UB

Anyway, yes, I know that it's possible to use values outside of their lifetime when you use pointers/references as opposed to smart pointers/weakrefs.

If this is a problem for you, you can use smart pointers, weakrefs, and be sure to assert in important places

It's not my really my problem tho if you're unable to use a low-level programming language I guess

You're just not the right person for the job then, right? 😄

[leaden vessel]

In any case, even supposing that following all the safe C++ practices helps, you still have to deal with the old unsafe legacy of the language, the compiler unable to check as much about the safety of your code as Rust can, and safety not being the default, making it harder to write code that is safe from the get go.
Here's another article about C++'s problems in that regard
https://www.thecodedmessage.com/posts/c++-papercuts/

[leaden vessel]

I'm perfectly capable of using C and C++ to do the job, I'll just have to deal with having potentially more memory safety issues that I'll have to fix later on

[quick bough]

there is no unsafe legacy. I use very little third-party code, and what I do use, is heavily tested

[leaden vessel]

Huh? There definitely is unsafe legacy in C++. Just because you make everything from scratch doesn't mean that's false

Heck, making everything from scratch means ditching the entire ecosystem. Not everyone can afford doing that, it's nice when you can though

But yeah sure I'm just not fit to be a programmer, totally not just a design and legacy issue

It's always the same argument, "you're just not good enough", meanwhile the skill issue just doesn't exist in another better ecosystem that doesn't make any significant trade-off in that regard

[quick bough]

I didn't mean to offend you

[leaden vessel]

Oh you totally did mean to. That's like the only thing "you're just not the right person for the job" is ever meant to do actually. You don't need to deny that :p
I'm not personally offended, I know my skills, I won't let myself feel insulted by statements like that. But I do feel somewhat pissed off by this completely backwards logic, which is why I engage in these kinds of discussions even though it's probably not good for me in the first place

[quick bough]

It's a joke on "right tool for the job"

Anyway, you really just seem to be arguing from a place of second-hand knowledge, hence it's really more a problem for you personally not with the language as a whole imo

[leaden vessel]

Heck, suppose I'm not the right person for the job. I've seen enough experienced C++ developers go to Rust and eventually actually understanding its benefits. I don't need to be the one saying stuff like that, I just don't like misleading statements or ones that are just plain wrong

[quick bough]

because clearly we are capable of writing complex software in C++ without a memory issue every 2 seconds

[leaden vessel]

Certainly not every 2 seconds, more like 1 out of 1000 lines of code for Google devs

Maybe a bit less than that if you follow modern practices even more than they do, idk I haven't checked, I just know they have this problem

I stopped doing C++ 5 years ago anyways

[quick bough]

I would be hardpressed to come up with a memory issue that isn't contrived

but yeah I can only speak on my own experience

which is in pure C++ codebases with a very good foundation

[leaden vessel]

I will concede that essentially making your own ecosystem and having better foundations outside of C++'s old legacy does help a lot
That's what raddebugger and Handmade Hero do, though in C and not C++

[quick bough]

and I do have a little secret trick: if there is an access violation, we can just catch it and continue execution after changing the registers a little 😉

[quick bough]

see, I don't understand why they'd pick C over C++ tho

not having to worry about having to free resources is a huge gain alone

anyway, I do have to "lock in" now... I'm trying to interact with a HID device and need to actually get to parsing the damn data

I'll let you know if I run into any memory safety problems 😛

[leaden vessel]

They don't use malloc and free at all

https://www.youtube.com/watch?v=TZ5a3gCCZYo

[quick bough]

I don't mean literal free, you might need to dtor some resources or whatever...

but also no matter how you allocate memory, you would still need to free that memory in some way

(unless you're a GC, then you can just leak it and call it a strategy, I guess)

[leaden vessel]

They don't use new/delete either, the point is they don't use RAII

[quick bough]

well yeah of course they don't use RAII, it's C after all 😄

[leaden vessel]

They use arenas instead, for group allocations

[quick bough]

Cool

The thing that RAII does is that it frees resources once they're out of scope

It's basically what you're used to with Rust

[leaden vessel]

yes

I know xd

[quick bough]

Surely they have some resources that do need deallocation/freeing/whatever

and because it's C, they have to do that manually

that's gotta be quite the pain, I imagine

[leaden vessel]

The argument goes that when you use arenas and allocate things in groups, lifetimes and resource management become trivial

[quick bough]

RAII is very useful for abstraction imo, so much less to think about, because you can just trust it all gets dtored/freed/whatever

[leaden vessel]

Can you give me an example?

[quick bough]

🙄

[leaden vessel]

That way I can contrast it with an arena equivalent

[quick bough]

yes, this is why I like my tree structures

[leaden vessel]

Tree structures are just nodes referencing each other in the same arena in their paradigm

[quick bough]

arenas don't avoid needing to free tho

[leaden vessel]

Yeah, you just arena.free_all() and you don't have to worry about freeing nodes since everything gets freed

It's also just writing to one pointer

[quick bough]

that's a huge yikes

this only works if a ton of guarantees are given by the nodes

[leaden vessel]

The nodes just need to be in the same arena

Nothing else is needed

[quick bough]

no, if the nodes own any resource, e.g. a file handle, you can't just free this node

you need to let this node know that it needs to relinquish its resources

[leaden vessel]

You're still thinking in RAII

[quick bough]

and then you can free it

[leaden vessel]

You don't have objects that just own resources in this paradigm

Resource management is done differently

[quick bough]

exactly, that's a guarantee, that the node has to give you.

if the node doesn't give this guarantee, you have a leak

[quiet root]

i hate to break this up but this has moved far away from wooting related dev stuff to just general discussions about languages

[leaden vessel]

ah shit that's true

[quick bough]

it do be like that sometimes tho

[leaden vessel]

At this point I'll just say you can ask them how they do it on their server and I'll stop talking about this

I guess #🔌│tech would've been more appropriate

or something

[quiet root]

yes

tech is the best for very general discussions like this that just relate to well tech

[quick bough]

people are already having a conversation in there 🙂

Honestly, I'm kinda impressed wooting's analogue input is so consistent

some of these I'm trying have some weird input issues where sometimes they just report as 0.2mm down

or don't report all the way down to 4mm instead stopping at like 3.8

[floral crane]

I haven't even touched the analog sdk or anything that uses it other than the built-in effects and gamepad lol

[quick bough]

I need to add support for more analogue keyboard vendors

Really feels like a pointless battle to fight to begin with at this point

but maybe these things just need time

[floral crane]

can you use the rgb sdk and analog sdk at the same time from the same process? like can I read analog state and send rgb updates? If I can I could use the analog sdk to read the fn key

[quick bough]

yes

Another one for the pile

Despite the inaccuracies with some of the keys, tho, I love the clickyness, reminds me of the original Wooting Two

[vestal stump]

can i somehow do an profile to main so when i start my pc is autoselects this one profile?

[kind ginkgo]

Not the right place to ask this but make sure you've got the profile you want to start with at the top

It'll always start with the first one

[vestal stump]

ah thank you

[cyan saddle]

/** @brief Reset all colors on keyboard to the original colors.

This function will restore all the colours to the colours that were originally
on the keyboard.

@ingroup API

@returns
Return true (1) if the keyboard is reset
*/
WOOTINGRGBSDK_API bool wooting_rgb_reset_rgb(void);

/** @brief Reset all colors on keyboard to the original colors.

This function will restore all the colours to the colours that were originally
on the keyboard and closes the keyboard handle. This function should always be
called when you close the application.

@ingroup API

@returns
Return true (1) if the keyboard is reset
*/
WOOTINGRGBSDK_API bool wooting_rgb_close(void);
WOOTINGRGBSDK_API bool wooting_rgb_reset(void);

What's going on with this undocumented function?

[quick bough]

Given the way it's spaced, I'd say it's an alias

Quick look at the source confirms that

[weary fern]

Hello, Can I try to add wooting SDK to my own pcb and try to make it work with wootility?

[quick bough]

Uhh, no, the Wootility is exclusively for Wooting keyboards.

[quick bough]

The most you could do as a keyboard vendor is provide a plugin for the Wooting Analog SDK so any applications using it can also be used with your keyboard, similar to what I did here.

[quick bough]

This strategy kinda worked, now I have customers reporting instances of digital input and asking for them to use analogue input, instead. 😄
Only downside: More work for me.

Which is funny because controller users never noticed or cared. I think with an analogue keyboard, tho, you expect far more of an impact because every. single. input. is analogue.

[cyan saddle]

Weird.

[quick bough]

Why weird?

[cyan saddle]

You can have that be exported in the library itself without exposing it in the headers. It seems like something was renamed.

[balmy iron]

I imagine it was done like that to not break any existing code, whatever that might be

[quick bough]

or they just wanted to provide this alias 😄

[bright valley]

What exactly does tacheon mode do?

Is it overclocking the processer on it? or

[quick bough]

Tachyon mode optimizes your keyboard for input speed. When enabled, every keypress responds as fast as possible. The RGB effects are disabled and the brightness is soft capped for an optimal result. Analog stability might be affected.

[bright valley]

Why don't they just use smt like embassy, to priotize input speed all the time?

and maintain polling rate etc.?

[quick bough]

embassy?

[bright valley]

rust crate

[quick bough]

I don't see how using Rust on the firmware would help anything in regards to speed

They very very likely use C for the firmware 😄

[bright valley]

Ig, I was under the impression that tachyon mode limited rgb to maintain polling rate etc.

and was wondering why they don't just do that by default?

[quick bough]

Wooting will not tell you their trade secrets obviously

[bright valley]

obviously, just taking a guess ig as to what goes on with tachyon mode lol

[quick bough]

At least part of tachyon mode, tho, is that the keyboard/firmware does less non-input stuff

So, RGB effects would obviously be wasting CPU cycles on "irrlevant" stuff, hence it gets disabled

[bright valley]

That's kind of what i was talking about with embassy, you could have the tasks like input stuff be above everything else to priortize that

[quick bough]

Reallistically, will you notice 1-2ms of latency? Absolutely not. 😄

[bright valley]

and help the cpu cycles

[bright valley]

lmao

[quick bough]

If I had a cent for every time a Rust programmer talked like they knew programming, it would still not make up for the frustration I feel. I should get at least $10 instead of only a cent.

Not like I know much about programming a keyboard's firmware either, but the simple truth is that your CPU either does something or it doesn't

And doing something takes time

However little time that may be

Also just Rust alone is generally considered a slower language (to C and C++, albeit less so) because it does also have some more runtime overhead

Again, realistically, does this time overhead matter?

If you're Wooting and there's people with solenoids and whatnot kind of equipment absolutely nit-picking your product, then maybe it does...

[bright valley]

Hmmmmm? lmao I don't really use Rust, much more of a cpp kind of person. But have used embassy in the past for drones

[bright valley]

idrc if it matters or not, I was just trying to see what all tachyon mode does, and why they do it in that specific way

[quick bough]

You're just doing too much high-level programming

[bright valley]

Probably

[quick bough]

Try writing a scheduler yourself, it's actually not that hard

[bright valley]

The only embedded experience I have is working with drones, etc. So I could be pretty wrong in what I assumed

[quick bough]

In my limited experience with Wooting's software design, there's like a 50/50 between "marketing" and "actual usability concerns"

So, I do assume the reason why RGB is limited is obviously for best possible functionality

[bright valley]

lol, ig I just saw tachyon mode as a way to avoid buying a better processer

[quick bough]

but the reason why this feature exists to begin with... it's just for that kind of person

hence why it's not the default

[bright valley]

yea

[bright valley]

Oh, the thing i was talking about already has been done before lmao
https://github.com/HaoboGu/rmk?tab=readme-ov-file

lmao discord ;-;

[quick bough]

So true

There is something kind of hilarious to me about people releasing a keyboard on Github

Like yeah bro just get this PCB printed and put it in a case

Anyway, I was familiar with the QMK project (and the related VIA protocol/interface/whatever), and this seems to be kind of a Rust version of QMK?

My disappointment on that front is that QMK is only focused on digital input which is almost too boring to even pay attention to

Like wow just read some I/O pins then report their values over USB

[bright valley]

its a full rewrite with almost no inspiration

just that they named it RMK to get more views

because theres ZMK a rewrite of it with bluetooth support imo a propietary 2.4ghz is much better than BLE support

[bright valley]

lol QMK does have HE support now

Well at least they should, if keychron would stop violating gpl 3.0 ;-;

[quick bough]

Yeah... that's the part that I'm disappointed about.

I'm a disappointed owner of a Keychron Q1 HE

Honestly, Drunkdeer keyboards are much better and they're not violating GPL 😄

but, still, no open-source firmware for analogue keyboards...

[bright valley]

I was genuinely about to buy one, but I looked at it and was like huh???? Where do the source code be

I really wanted a knob on my keyboard :(

lmao its the only reason i didn't buy a zoom75

[quick bough]

yeah ikr, it's all over their store page

even on the damn package for it

[bright valley]

i got the QMK devs to harass them for it

and now they say they'll give it in october

but like why october? why not now????

what changes in october lmao

[quick bough]

I think they're embarrased by how bad their code is

[bright valley]

Like thats not how GPL works my friend

[quick bough]

Well, what are laws if not friendly notes? 😄

If the QMK folks aren't gonna sue before September...

[bright valley]

Idk why they don't send cease and desist, theyre like its a cat and dog chase if we have to do it with everyone

IMO do it with at least keychron ;-;

like sure qwertykeys will just avoid them etc.

but keychron? you can really get a lot of their customers to go away by being vocal about it

[quick bough]

Why is there a 'T' in the keychron icon tho

[bright valley]

¯_(ツ)_/¯

why not

[quick bough]

All the hype around snap tap is insane

I don't envy the QMK maintainers

[bright valley]

all the random programs that implement scd for free arek inda funny

[quick bough]

Not least because they ship 300 MB executables and think it's okay

[bright valley]

what'd they do to you?

[bright valley]

lmao why is it 300mb?

[quick bough]

because they think .NET 6 is greater than .NET 4.5

(now they have to ship the entire runtime because it's not shipped with the OS)

[quick bough]

"don't envy" means they're in a bad position

like everyone just wants to have the snap tap

but obviously filthy digital keyboard users will only be able to get a nullbind

which, from what I hear, is decent enough

but not like there weren't already scripts for this like 20 years ago

except now everyone feels like because the cheat is on the hardware it's suddenly okay

[strong siren]

isnt snap tap a purely digital feature?

[quick bough]

not really?

[bright valley]

They already have SOCD implemented

it's just not out in the main thing yet

[strong siren]

last input priority socd handling doesn't really require analog input

the way razer does it, they don't even use rapid trigger to their advantage

[bright valley]

because razer and wooting did it, its no longer scripting

[strong siren]

at that point it's just adjustable actuation point last input priority socd handling

that's a mouthful lol

rappy snappy is analog specific but it's much less noticeable from what i read here

[quick bough]

well, I think what razer did is slightly more complex than last input priority

[bright valley]

is it?

[strong siren]

in what way?

[quick bough]

but yeah, you are right, in principle you can do this SOCD resolution digitally

[strong siren]

i might be misunderstanding what they did so I'm genuinely asking

[quick bough]

Who am I to reveal Razer's trade secrets now? (I just sniffed the HID traffic a bit, lol.)

[strong siren]

https://github.com/cafali/SnapKey
this sounds right up your alley lol