Fix compilation for ballista in stand-alone mode

[Igosuki]

Which issue does this PR close?

Closes #1000

Rationale for this change

API changes broke compilation and execution of the standalone ballista client.

What changes are included in this PR?

Compilation fixes, and another fix is required for tests to pass

Are there any user-facing changes?

None

[Igosuki]

Problem here https://github.com/apache/arrow-datafusion/blob/master/ballista/rust/executor/src/standalone.rs#L55
By the time df collects test results the tonic server is unreachable, I tried to make it run ad-vitam but then the tests never end.

[alamb]

I verified that the cargo check --features=standalone compiles on this branch, but does not on master

When we fix the ballista test, I recommend we start running the standalone client tests as part of CI so they aren't broken again

[alamb]

This seems like an improvement to me, so merging it in. We can always improve things more in the future

[alamb]

Thanks again @Igosuki

[Igosuki]

So how do we fix the ballista test though ? I don't understand why the tonic server is unreachable since even if the thread gets detached because the tokio JoinHandle gets dropped it should still be running right ?

[alamb]

I wonder which part of the not is not parsing.

Sorry @Igosuki I am not super familiar with the ballista codebase yet. What test are you referring to (aka what command are you running locally that is failing)?

With that information I can file a ticket for the problem and probably find someone in the community who can help fix it

[Igosuki]

@alamb This one https://github.com/apache/arrow-datafusion/blob/master/ballista/rust/client/src/context.rs#L300

[alamb]

Filed #1020 to track test failing