Hi,
In the context of device image engineering, I’m wondering how the recently introduced system config interface differs from specify connection policy in a gadget snap.
Historically, we moved away from using the gadget based connections policy because we found occasional issues when provisioning devices, where the policy wasn’t respected. Instead we moved to store based policy only, which proved very reliable.
Although store based policy seems more robust, it is more painful to change because it’s not self service for privileged or complex interfaces.
On the surface, it seems like the interfaceconfig could be a good option, especially as it can be updated easily after device provisioning [ although it doesn’t get enforced until the target snap refreshes next ]. But, I’m wondering if there are any other technical differences between that and gadget snap policy to be aware of, or any other limitations that aren’t obvious.
Any experience or insights would be appreciated.
Thanks!
Cheers,
Just