Network improvements

[2bndy5]

This ports changes from the following recent PRs to the pure python codebase:

• nRF24/RF24Network#227
• nRF24/RF24Mesh#238
• nRF24/RF24Mesh#240
• nRF24/RF24Mesh#244
• nRF24/RF24#955
• nRF24/RF24Mesh#250

[codecov[bot]]

Codecov Report

Attention: Patch coverage is 25.49020% with 38 lines in your changes missing coverage. Please review.

Project coverage is 86.27%. Comparing base (fc69216) to head (492de7d).

Files	Patch %	Lines
circuitpython_nrf24l01/rf24_mesh.py	18.18%	36 Missing :warning:
circuitpython_nrf24l01/rf24.py	0.00%	2 Missing :warning:

Additional details and impacted files

```diff @@ Coverage Diff @@ ## master #59 +/- ## ========================================== - Coverage 86.68% 86.27% -0.42% ========================================== Files 16 16 Lines 2463 2484 +21 ========================================== + Hits 2135 2143 +8 - Misses 328 341 +13 ```

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.

[2bndy5]

I think the bug fixed in nRF24/RF24Network#229 is specific to the C++ implementation. Looking at the _write_to_pipe(), I forgot how much code gulf I played to speed things up.

There is no RF24Network.networkFlags here because

• message fragmentation is done directly in _write_to_pipe() (making FLAG_FAST_FRAG obsolete)
• FLAG_NO_PULL is implemented as _parenthood, even though there is no RF24Gateway/Ethernet implementation in this CirPy port

Furthermore, the network layers here don''t use a writefast() equivalent. Instead, I use a regular blocking write() (RF24.send()) and call RF24Network._tx_standby() which uses RF24.resend() similarly to how C++ RF24::txStandBy() uses RF24::reUseTx() but in a blocking manner.