If you attempt to call peripheral.connect from the callback of set_callback_on_scan_found or set_callback_on_scan_updated, self.peripheral.state will be forever stuck in CBPeripheralStateConnecting until the WAIT_UNTIL_FALSE_WITH_TIMEOUT(self, _task.pending, 5.0) times out.
This is because set_callback_on_scan_found callbacks are called from delegate_did_discover_peripheral which is called from a CBCentralManager delegate callback, which is run on the CBCentralManager's queue. I'd guess the CBCentralManager does updates to all of its data structures on that queue (this is generally how Apple APIs handle thread safety), so by blocking the queue (by sleeping in connect) you prevent the code that updates peripheral.state from running.
This probably applies to any other callbacks as well (e.g. the notify callback).
If you attempt to call
peripheral.connectfrom the callback ofset_callback_on_scan_foundorset_callback_on_scan_updated,self.peripheral.statewill be forever stuck inCBPeripheralStateConnectinguntil theWAIT_UNTIL_FALSE_WITH_TIMEOUT(self, _task.pending, 5.0)times out.This is because
set_callback_on_scan_foundcallbacks are called fromdelegate_did_discover_peripheralwhich is called from a CBCentralManager delegate callback, which is run on the CBCentralManager's queue. I'd guess the CBCentralManager does updates to all of its data structures on that queue (this is generally how Apple APIs handle thread safety), so by blocking the queue (by sleeping inconnect) you prevent the code that updatesperipheral.statefrom running.This probably applies to any other callbacks as well (e.g. the notify callback).