Replies: 4 comments 3 replies
-
Was thinking about it further and I think that we'd probably want one test per http code that could be returned. So, we'd have a success case for We'd probably have something similar for non-200 status codes, but I think we can be less specific there (by saying non-200) except in cases where a specific code is expected. |
Beta Was this translation helpful? Give feedback.
-
I think @odata.id should be a MAY as well:
Assuming of course that metadata=minimal is still the default unless otherwise specified. OData 4.0 stated that it is the default, but the text was removed from OData 4.01
|
Beta Was this translation helpful? Give feedback.
-
On testing as a whole, I think the above is a great starting point. It might have to change after we start testing servers, but it's going to be very hard to know that until we do start the testing. So, I am in favour of going forward with the testing rules above. |
Beta Was this translation helpful? Give feedback.
-
If the writeup above is still current/relevant, I have a concern:
Supporting DELETE should be a MAY instead of a MUST for a couple of reasons:
|
Beta Was this translation helpful? Give feedback.
-
Here is the initial set of BDD acceptance testing rules (Gherkin Reference) reviewed on the June 17, 2021 Certification Subgroup call:
Feel free to make suggestions, change things entirely, ask questions, clean things up, or add things. We'll discuss this further in the Transport and Certification Groups in March-April 2021.
cc: @EnFinlay @dgmyrek @SergioDelRioT4Bi @psftc @RESOStandards/employees
Beta Was this translation helpful? Give feedback.
All reactions