Start on Bulk API testset

[enikao]

No description provided.

[dslmeinte]

Can you add the description of this directory and file structure as a README somewhere?

[enikao]

Can you add the description of this directory and file structure as a README somewhere?

Will do

[enikao]

I'm not sure whether we want one file per test case listing all steps, or one file per step with the directory being the test case.

Pro file per test:

• Easier to execute, no file system shenanigans

Pro file per step:

• Easier to navigate, read, write

Con storing tests themselves as model:

• We would need to package the serialization chunk as property, as we want to test invalid chunks

[joswarmer]

I foresee a maintenance problem with embedding chunks in the JSON test files.

We are using different tools (Freon, MPS, Java/Kotlin, ...) to read and write chunks, because we don't want to write bare JSON. These tools typically read/write chunks from a file. It is going to be hard to read/write the chunks embedded in the test files using these tools.

Maybe we should put the chunks in separate files and use the $variable syntax or something similar as values for chunks referring to a file containing a chunk.

[dslmeinte]

I think the Java lib should mention the LionCore Built-ins as used language.

[enikao]

Maybe we should put the chunks in separate files and use the $variable syntax or something similar as values for chunks referring to a file containing a chunk.

How does this solve the problem that we have to refer to node ids we've got from a previous request?

[joswarmer]

Maybe we should put the chunks in separate files and use the $variable syntax or something similar as values for chunks referring to a file containing a chunk.

How does this solve the problem that we have to refer to node ids we've got from a previous request?

It doesn't, it is intended to make creating and changing the chunks in the testcases easier.