Fix reference definition range validation

[ramonski]

Description of the issue/feature this PR addresses

This PR fixes the field validation for the result, min and max values of the reference definition ranges.

Current behavior before PR

• min value can be above max value
• min/max values are kept when result changes, e.g. it is possible to change the result above the max value or below the min value afterwards.

Desired behavior after PR is merged

• min value is validated to be below max value
• min/max values are shifted when the result changes

--
I confirm I have tested this PR thoroughly and coded it according to PEP8
and Plone's Python styleguide standards.

[xispa]

This referencevalues_validator is still called at https://github.com/senaite/senaite.core/blob/fix-referencedefinition-values-validation/src/bika/lims/content/referencesample.py#L173

[ramonski]

Good catch, thanks for searching over the codebase once again! Fixed in b75d84e

[ramonski]

Do you think it is worth to make a migration step for this?
E.g. if someone defined a higher max than min value?

[xispa]

Do you think it is worth to make a migration step for this? E.g. if someone defined a higher max than min value?

Good point!. Yes, I think is worth. I don't expect instances to have a lot of Reference Definitions, so the upgrade step would not take too much to complete.

[xispa]

Note we can still have min > max values in reference definition when user inputs a negative percentage:

We have some options here:

1. Move this code after line 234
2. Do s_err=abs(float(s_err)) before line 232. In such case, we might need to also do "error": str(s_err) at line 236.
3. Rather consider negative percentage as a feature and leave it as it is

What do you think?

[xispa]

Note we can still have min > max values in reference definition when user inputs a negative percentage:
[...]

With c23508d we do the abs, so min above max are no longer possible.