Semantic E2E Tests

[tzaffi]

Problem

When creating new PyTeal constructs (e.g. #198, #196), or modifying existing ones, there are currently no tests to validate that resulting constructs are semantically correct. We have end to end tests that validate that generated TEAL is as expected, but it would be even better to validate that the generated TEAL has the intended meaning.

Solution

Introduce Dry-run capabilities into the e2e tests so that generated TEAL can be run and results be validated.

Dependencies

None

Urgency

Medium. I think this fills in a gap in the current PyTEAL development cycle.

[tzaffi]

Added #179 as a blocker, as the current PR #249 as it currently stands, requires a change to @Subroutine's interface, and this could conflict with what ABI ends up doing (and which has a higher priority)

[tzaffi]

this is merged in