Use actual bytesarray/memoryview instead of bytes concatenation

Bytes concatenation such as this is extremely wasteful and is never a good idea.

https://github.com/fluent/fluent-logger-python/blob/master/fluent/sender.py#L137
```python
# buffering
if self.pendings:
self.pendings += bytes_
bytes_ = self.pendings
```

Python has mutable buffers that should be used to ensure that the objects aren't created needlessly putting pressure on GC and wasting CPU with double-copy concatenation.

Start in fluent/sender.py at line 137 and trace the buffering path around self.pendings. Read the surrounding sender implementation to compare the proposed bytearray or memoryview approach with the current concatenation. Done means buffered sends no longer repeatedly concatenate bytes while preserving the sender's existing behavior.

Scritto dal modello di indicizzazione a partire dal testo della issue.