noports
noports copied to clipboard
at_activate in C
Dependent on atsign-foundation/at_c#176
### Tasks
- [ ] https://github.com/atsign-foundation/at_c/issues/176
- [ ] atsign-foundation/at_c#74
Where would this project live @XavierChanth ?
Would it be in examples/ ?
I'm guessing here, as @XavierChanth hasn't explained why...
at_activate is a crucial component of the customer journey, and if their device needs C sshnpd then it likely also needs a smaller and more portable at_activate.
So not an example really, but also not part of the SDK. Maybe we need a utils directory, or maybe we need to put it in the noports repo alongside the daemon.
One thing we could also consider is making an entirely new at_tools repo made in C, like at_tools_c that would house generic Atsign tools like repl, at_talk, atcat & attee (https://github.com/atsign-foundation/at_demos/issues/309), and also our at_activate.
repl and at_talk currently live in our examples/ directory.
The primary object for creating at_activate is to make it easier to do APKAM / atSign onboarding with the C daemon. I am hesitant to put it in examples, but maybe a bin/ directory at the root of this repo. repl & at_talk can be moved into this new directory.
This is now something that needs to float to the top of the ToDo's.. It will enable a number of customers to automate sshnpd and its keys on IoT devices without logging into the hardware.
MVP at_activate with APKAM creation of atKeys
I will take this up this sprint @cconstab
APKAM functionality is the target.. TY
We would like to move this ticket to NoPorts. We would like to have at_activate be within the NoPorts repo to make things simple for our CI when we package NoPorts zips and binaries
Would it be possible to change this ticket's repo? I don't know how to do that
Carrying forward to PR97 to implement key/keypair generation and exchange these keys with the at_server.
@JeremyTubongbanua, @XavierChanth and I have completed the implementation of dependencies required to implement at_activate. But there is one bug which was interrupting my progress (MbedTLS read returns a connection timed out err), @JeremyTubongbanua is going to take a look at this bug today as I'm not very familiar with MbedTLS and it's eating up a lot of my time.
In the meantime, I have started working on at_activate itself. Current ETA to complete this would be Tuesday (at the latest).
Blocked by https://github.com/atsign-foundation/at_c/issues/426
https://github.com/atsign-foundation/at_c/issues/426 has been resolved. See ticket for explanation.
Hows this ticket progressing.. looks like it is close but hard to tell .. Itching to test :-)
All the implementation is complete @cconstab. Trying to resolve all the obvious bugs. You should have a version to test in a day or two.
@purnimavenkatasubbu found a bug in an atserver that was activated with C atactivate. The server throws a publickey not found error when updating data. Looking into this.
Pending items:
- above mentioned bug
- unit tests (if there are any unit testable elements)
The at_activate PR has been merged. The implementation to create enrollments with OTP is being tracked through https://github.com/atsign-foundation/at_c/issues/451.
Closing this ticket as the implementation to create new enrollments using OTPs has been merged to trunk through: https://github.com/atsign-foundation/at_c/issues/451