Inconsistent gap treatment in ring buffer

[cwasicki]

What happened?

If there is no data in the ring buffer (e.g. on start up) or no valid data (e.g. only None values), there is inconsistent behavior for gaps and __len__, which relies on gaps.

What did you expect instead?

Example on startup (expected 1 gap over the full window):

frequenz-sdk-python/tests/timeseries/test_ringbuffer.py

Line 211 in 01b8d9d

assert len(buffer.gaps) == 0

Example for no valid data (expected length of zero):

frequenz-sdk-python/tests/timeseries/test_ringbuffer.py

Line 252 in 01b8d9d

assert len(buffer) == 4 # bug: should be 0 (whole range gap)

Affected version(s)

v0.x.x

Affected part(s)

Data pipeline (part:data-pipeline)

Extra information

No response

[cwasicki]

Adding it to rc2 milestone because it would ideally be addressed for v1.0 release. But could also be addressed later since it's a bug fix which affects a very special case only.