Fix tuples handled incorrectly in decode_function_input

[davesque]

What was wrong?

Fixes #1409 .

How was it fixed?

Canonical tuple type strings weren't being properly generated in decode_function_input.

Cute Animal Picture

[davesque]

This should fix the immediate issue. Going to see if the tests pass and also look for any other places where type strings aren't being generated correctly.

[davesque]

I think this is good to merge with approvals.

[pipermerriam]

Any chance there's a test that could have caught this?

[davesque]

@pipermerriam Yeah, good point. Let me add one.

[kclowes]

Ack, that parity failure is a flaky test. You should be able to re-run and get a pass

[davesque]

@pipermerriam I thought about adding more complex tests, but I think that functionality is already covered by the tests that were added when tuple support was merged. It seems like the failing example from #1409 is good enough to ensure that get_abi_input_types is being called, which is what we should be testing for.

Not sure what that failure's all about. Seems unrelated to this change.

[carver]

I'm not deep in this right now, but it looks fine to me! 👍

[ankitchiplunkar]

thanks a lot @davesque

[davesque]

@ankitchiplunkar No prob, cheers!