Wolfram Richter

so I personally do think that GPIOD is the better solution, and I have been running the implementation successfully for the last ~4 months. But it hasn't seen much testing...

This is implemented in #525

> May I ask why _active_low_ ist set true in _MeterS0.cpp_ when using _libgpiod_? I did not look at the sysfs implementation since that was a non-starter on Fedora. When...

Legacy makes a compelling case :-) My hardware lives on a breadboard, so I can somewhat simply rewire towards a pulldown resistor. I believe that this should be configurable and...

@trabant-rgb Agree with your comments. I suppose I need to reach out to the developers mailing list to get help; the GitHub discussion does not create much interaction with the...

Thanks to @r00t- 's merge the unit tests are now completing successfully: https://github.com/wrichter/vzlogger/actions/runs/3943901824/jobs/6749257157

> would also suggest to take my change to add some verbosity to cmake. Thanks, did that. > code-wise i'm unsure what to make of this. i haven't had time...

> I was using the `gpiod` branch very successfully with an electronic (Cyble) sensor using the `high_wait` option and no `debounce_delay`. Now I moved and have a different gas meter...

> But then, another rising_edge causes the state machine to go to 3/high_wait and further back to 1/high, following by a wrong reading being counted, even if the state_machine was...

Ok this might be minor semantics, but what you seem to want is a LOW_WAIT state that enforces a minimum duration low signal between two highs. Debounce is really just...