Allow getting an uninitialized DelegateParameter with offset or scale #3662
Conversation
…one; added two tests
haroldmeerwaldt
changed the title
|
@astafan8 I put the tests in test_parameter_scale_offset.py as you asked. A good place would also be test_delegate_parameter.py. As I was scrolling through it, I found that the previous behaviour may actually be intended: What do you think about the other location, and the change of expected behaviour? |
true, but i don't think that matters much. leave this decision up to your preference :)
hmmm..... that's a very good question! glad that we have these tests so that we think about this. However, i think this test is only capturing the current behavior as opposed to making it "desired". I don't think it makes sense that get() and snapshot() rasie a TypeError with the scale/offset set and ready while the underlying parameter's value is None - that's where the original issue is coming from, right? that this behavior does not bring value. we can ask for a third opinion - @jenshnielsen what is your take on this? |
|
I think its fine to change the test to reflect this new behavior. Just make sure that we document this by adding a changelog fragment |
Codecov Report
@@ Coverage Diff @@
## master #3662 +/- ##
=======================================
Coverage 65.74% 65.74%
=======================================
Files 225 225
Lines 30524 30524
=======================================
Hits 20068 20068
Misses 10456 10456 |
|
bors merge |
_from_raw_value_to_value)Fixes #3653