I would like to propose two improvements for this plugin.
both of described things are coming from personal experience setting up the thermostat and homebridge on rpi2.
sometimes homebridge can't connect to the device and then it can't reconnect until I restart the homebridge. It means that if connection to the device timed out - then there is possibility that device went off the grid - i.e., it shouldn't be discoverable - i.e. we need to re-run discover
sometimes I start homebridge too early - much earlier than the thermostat boots up and finishes all it's adoptions. in this case homebridge won't pick up this thermostat when it will be ready - it will just stay as non-working device. It means that if discover failed in constructor - then we shouldn't block all further work with this thermostat - we need to retry discover on next request
I would like to propose two improvements for this plugin.
both of described things are coming from personal experience setting up the thermostat and homebridge on rpi2.
sometimes homebridge can't connect to the device and then it can't reconnect until I restart the homebridge. It means that if connection to the device timed out - then there is possibility that device went off the grid - i.e., it shouldn't be discoverable - i.e. we need to re-run discover
sometimes I start homebridge too early - much earlier than the thermostat boots up and finishes all it's adoptions. in this case homebridge won't pick up this thermostat when it will be ready - it will just stay as non-working device. It means that if discover failed in constructor - then we shouldn't block all further work with this thermostat - we need to retry discover on next request