pymeasure: check_errors usage inconsistent

Our Instrument.check_errors() implementation returns a list of received errors (as its docstring says).

Interestingly though, where we use them, e.g. in control’s fget, https://github.com/pymeasure/pymeasure/blob/ebf6df97e229c0e916192a495ade65053b234ac3/pymeasure/instruments/instrument.py#L361, we simply discard those return values without doing anything!

I think where we use it we should do something with the errors, most probably log them.

About this issue

Original URL
State: closed
Created 2 years ago
Comments: 24 (24 by maintainers)

Most upvoted comments

What do you think about a slight duplication of logging?

We could retain the current behaviour of default check_errors, i.e.

log errors
return list of errors

check_get/set_errors document that they (if overridden) shall

log received errors
return a list of encountered errors
Optionally may raise an Exception if a critical error is encountered

Property fget/fset, if they receive an error during checking, log that error + useful context (command). So we might get

…error: KeithleyXYZ: 235, baz is borked (from check_errors) …error: Error received after writing “POW 1”: “235, baz is borked” (from within property fget/fset)

This way, we ensure that

we get logging on error by default, even if the error does not happen within a property
we get additional useful context if this happens in a property
error checking always returns something that can be easily processed in code
we document that an instrument may raise an Exception if deemed appropriate

Thoughts?

bilderbuchi on Apr 28, 2023