Do not test serialization modifiers in a value-dependent manner

[mjacoby]

There are a range of test cases that test what's been referred to in the specification as "applicability of serialization modifiers" (see spec).

However, the meaning of this table of obviously not clear as there is not definition of "applicability". This has been discussed in the TF API admin-shell-io/aas-specs-api#327

As the specification is underdefined and there is no official statement how this is expected to work the related tests should be removed.

[bjoern-otto]

Thanks, we'll have a look into this.

[juileetikekar]

Hi @mjacoby, thanks for the input. The current behaviour around the applicability of the modifiers has been explained and added to the transparency list of the test-engine: Non-Applicability of Serialization Modifiers . And you are right, the behaviour is subject to change as per the GH Issue admin-shell-io/aas-specs-api#327

[bjoern-otto]

We decided to keep the tests as they currently are until the specification group has decided how to handle this appropriately.
In the meantime, we suggest to use filtering in case you do not need these tests:
https://github.com/admin-shell-io/aas-test-engines?tab=readme-ov-file#running-a-subset-of-the-tests

[mjacoby]

Are you sure this is a good idea as this will imply that standard-compliant implementations might fail the tests using default settings for at least until the next release of the specification which I think is not planned earlier than beginning of Q2/2026?

Or are there any concrete plans to resolve this in the specification earlier than that which I am not aware of?

[bjoern-otto]

@juileetikekar Maybe you can clarify?

From a technical point of view, you are right. It is more a thing of prioritization as we do not want to remove existing functionality without an replacement.

[bjoern-otto]

(re-openend as a future reminder once the spec is updated accordingly)