Zwift Click handled

[cagnulein]

A very raw Zwift Click implementation. It worked
Feel free to clean the code!

[ajchellew]

Thanks for this, very interesting and given I don't have a Click, it gives me some insights

Confirms my guess in #2 that BrevitDeviceType BC1 is the Click, and all the assertions that this protocol is not just for the Play Controllers.

• Would it be possible to provide a few lines from:

  Logger.d("Click Button Press ${data.toHexString()}") ?
  I'm curious to see what Protocol buffer type this is, at least currently I haven't managed to spot it from what I know.

• Does the Click connect to Zwift and Zwift application relay to the Hub, this is my other assumption that I can't prove?

I will try to change the code so that the Click and Play are handled separately (just a bit OCD unneeded-ness)

[cagnulein]

Confirms my guess in #2 that BrevitDeviceType BC1 is the Click, and all the assertions that this protocol is not just for the Play Controllers.

Exactly, it was my thought too

• Would it be possible to provide a few lines from:

  Logger.d("Click Button Press ${data.toHexString()}") ?

Sure

I'm curious to see what Protocol buffer type this is, at least currently I haven't managed to spot it from what I know.

• Does the Click connect to Zwift and Zwift application relay to the Hub, this is my other assumption that I can't prove?

I don't have the hub so I can't say this :) I bought this only to integrate into my qdomyos-zwift app https://github.com/cagnulein/qdomyos-zwift/

I will try to change the code so that the Click and Play are handled separately (just a bit OCD unneeded-ness)

Hehe sure! Go on!

I'm also porting this into swift.

I started doing this using openssl but it was so painful to debug

[ajchellew]

I bought this only to integrate into my qdomyos-zwift app https://github.com/cagnulein/qdomyos-zwift/

Its cool if this little boredom side project of mine has been any help. :)

I started doing this using openssl but it was so painful to debug

When I did the .NET port I initially tried to use the built in .NET crypto libraries, switched to BouncyCastle and got it working in no time.

[cagnulein]

Yeah unfortunately bouncecastle is not available for ios. I have to check about bouncecastlecpp Roberto Viola Software engineer and open source enthusiast http://robertoviola.cloud Il giorno sab 3 feb 2024 alle 12:40 ajchellew ***@***.***> ha scritto:

…

I bought this only to integrate into my qdomyos-zwift app https://github.com/cagnulein/qdomyos-zwift/ Its cool if this little boredom side project of mine has been any help. :) I started doing this using openssl but it was so painful to debug When I did the .NET port I initially tried to use the built in .NET crypto libraries, switched to BouncyCastle and got it working in no time. — Reply to this email directly, view it on GitHub <#3 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAALYWGDZZ67KAMZDA5BGTTYRYO2LAVCNFSM6AAAAABCUPQQBSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMRVGI4TIMJWGQ> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[cagnulein]

@ajchellew I just made an astonishing discovery.

if i send

" >> 52 69 64 65 4f 6e // handshakeStart"

the click sends the data without any encryption!

do you see the same on the zwift play device?

[cagnulein]

to answer to your request this is the async bytes returned

when no buttons are pressed

"37 08 01 10 01"

plus pressed

"37 08 00 10 01"

minus pressed

"37 08 01 10 00"

[ajchellew]

Wow, I'm certain the play controllers never worked without encryption. Maybe the click is a little less crazy since really who cares if that data is intercepted.

That said there's been at least 2 updates to the controller since I first did this, so maybe it's possible now.

I've got the Kickr Core and virtual shifting now, so I'll probably have a look over the weekend. It answered my question, the trainer and controllers talk via Zwift. They aren't paired, and it makes sense as you see what gear you are in on screen.

[cagnulein]

it works also on the play without encryption! you just need to reset them :) Roberto Viola Software engineer and open source enthusiast http://robertoviola.cloud Il giorno gio 8 feb 2024 alle 18:02 ajchellew ***@***.***> ha scritto:

…

Wow, I'm certain the play controllers never worked without encryption. Maybe the click is a little less crazy since really who cares if that data is intercepted. That said there's been at least 2 updates to the controller since I first did this, so maybe it's possible now. I've got the Kickr Core and virtual shifting now, so I'll probably have a look over the weekend. It answered my question, the trainer and controllers talk via Zwift. They aren't paired, and it makes sense as you see what gear you are in on screen. — Reply to this email directly, view it on GitHub <#3 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAALYWDROCTPLA3EG3254H3YSUAKTAVCNFSM6AAAAABCUPQQBSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMZUGU2TMMZRGE> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[ajchellew]

Kind of nice, makes the hardware easier to be reused. It's got to have changed in a new firmware, I'm not the only one that fought with the encryption. Makinolo would have had an easier time with his GTA mod

[cagnulein]

i don't know, maybe or maybe you simply didn't try this :) Roberto Viola Software engineer and open source enthusiast http://robertoviola.cloud Il giorno gio 8 feb 2024 alle 18:24 ajchellew ***@***.***> ha scritto:

…

Kind of nice, makes the hardware easier to be reused. It's got to have changed in a new firmware, I'm not the only one that fought with the encryption. Makinolo would have had an easier time with his GTA mod — Reply to this email directly, view it on GitHub <#3 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAALYWFWJF44U5PJ6IRN4LTYSUC5ZAVCNFSM6AAAAABCUPQQBSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMZUGYYDKMRYGM> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[ajchellew]

@cagnulein you are right, having looked at it I never thought to try RIDEON without the extra 2 bytes, I tried without the key after and obviously that didn't work.

I've just committed up a bit of a refactored version of this PR. I'm hoping it should just work for the Click but it might not. The data you provided helped add the decode in the long protobuf way. I've also added an ENCRYPTED flag that bypasses all that too.

[cagnulein]

Glad it worked! Il giorno sab 10 feb 2024 alle 17:59 ajchellew ***@***.***> ha scritto:

…

@cagnulein <https://github.com/cagnulein> you are right, having looked at it I never thought to try RIDEON without the extra 2 bytes, I tried without the key after and obviously that didn't work. I've just committed up a bit of a refactored version of this PR. I'm hoping it should just work for the Click but it might not. The data you provided helped add the decode in the long protobuf way. I've also added an ENCRYPTED flag that bypasses all that too. — Reply to this email directly, view it on GitHub <#3 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAALYWCDXFX35M5PJRYVSIDYS6RPNAVCNFSM6AAAAABCUPQQBSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMZXGA3DEOBUHE> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[cagnulein]

@ajchellew did you try the new firmware https://zwiftinsider.com/play-firmware-1-3-0/ ?

[cagnulein]

the new firmware works with qz :)

[ajchellew]

@cagnulein hadn't seen this, just updating them before I do a ride, I assume nothing changed in the BT protocol itself