Bluetooth indeed hops around in the 2.4 GHz band split in 100 kHz bands but Bluetooth has an algorithm to stay out channel that are occupied. For that to work, Bluetooth has off coarse to see the other "user" and if the channel is not used that often that can take time or will not work. But .. provided that Bluetooth stays out the band you are using for MySensors there is still a possible problem in that the nRF24L01+ (like almost all 2.4 receiving IC's) has a wide-band LNA followed by a mixer followed by filtering and more gain. The wide-band will see both the Bluetooth and the Signal Of Interest (SOI) and if the level of Bluetooth if much larger than the SOI things will also go wrong. So if you have Bluetooth at 50 cm, it will probably come in around -30 dBm, the nRF24L01+ can receive signals up to -85/-94 dBm and this kind of level differences can give problems, see also Table 7 of the nRF24L01+ datasheet.

The reason that the log is being filled is because despite not having the gateway ready message data can become available. I have been thinking of adding an extra switch in the software that if gateway ready is not received, but other data arrives to toggle the internal state and have the gateway marked ready. And when first data arrives ask for the version number. I do not have enough knowledge about the differences between the FTDI FT232RL (nano) and ATmega32U4 (pro micro) when they are ready to accept data to be send to/from the gateway. If the usb chip is not ready to accept and gateway ready is send, data is lost. But this is an assumption (Arduino's only have receive and not send buffers as far as i know), maybe some hardware guru can shine a light on it.

If you're talking about SerialGateway.ino.. then yes.