feat: Add new test suite for context merging

[chrfwow]

This PR

Adds gherkin tests to verify the merging order of contexts: https://openfeature.dev/specification/sections/evaluation-context/#requirement-323

See open-feature/flagd-testbed#13

[chrfwow]

Do you prefer the same line with different values each time, or a data table for the Given and Then parts of this scenario?

[aepfli]

A datatable could here be beneficial, can we tag on datatables, because we do have the same problem as above, what if not all levels are supported by the sdk, how can we limit this?

[chrfwow]

Apparently tags can only enable/disable a whole table, not individual lines in a table. So we would need a table for every permutation of possible implementations if I understood that correctly.
Alternatively I could add a tag for each of the levels, or at least for transaction and hook

[aepfli]

i am not sure tags work on steps, normally they only work on scenarios or examples. from the docs at https://cucumber.io/docs/cucumber/api#tags:

Tags can be placed above the following Gherkin elements:

    Feature
    Rule
    Scenario
    Scenario Outline
    Examples

[chrfwow]

Hmm, that's a problem. So how can we easily enable/disable these steps? Should we really duplicate all tests?

[aepfli]

As long as we only need to duplicate gherkin tests, and not have additional effort in code, i think that is a feasible solution. Gherkin scenarios are cheap, not a lot of complexity and easy to grasp, there is no code to understand nor bigger complexity to grasp.

[chrfwow]

I tried something new by dublicating most of the tests. I don't really like this solution that much, but we should now at least be able to support implementations with and without hooks and/or transactions. Refactoring this into a Datatable would make the code implementation a bit more complicated I think, so I left it for now as individual steps

[aepfli]

Thank you, i think those gherkin tests will make our live easier in the future. But we have to be a little bit more careful how we enable certain tests with tagging. not all sdks support the whole feature set, and we want them to use this files and limit the usage. We should think about how to properly tag certain tests, or move them in separate scenarios. eg. transaction context merging

[aepfli]

we might want to start adding tags here, not all sdks support eg. hooks or transaction - hence i think maybe adding a tag for context merges and the level would be good something

[chrfwow]

I agree, but here we also have the problem that tags can only be used on a whole table, not on individual lines inside a table

[aepfli]

A datatable could here be beneficial, can we tag on datatables, because we do have the same problem as above, what if not all levels are supported by the sdk, how can we limit this?

[aepfli]

looks good, we might want to have also one test with @Hooks @Transaction testing both together - i know it is another duplication, and to be honest, i do not have a hard feeling for or against them. Thank you for improving our gherkin tests.