Implement bitwise shift opcodes for Constantinople

[cburgdorf]

What was wrong?

As stated in #1104, the upcoming Constantinople fork adds three new opcodes for bitwise shifting operations. Namely SHL, SHR and SAR. This PR adds support for the SHL and SHR opcodes.

This does not include support for the SAR opcode yet which is expected to land in another PR. #1134

How was it fixed?

1. Created a new test_opcodes file and wrote a couple of tests for existing opcodes. This isn't finished but from the perspective of this PR it shouldn't matter and we have an open issue to not forget about testing other existing opcodes Create unit tests for Opcodes #1111

2. Implemented SHR and SHL logic in eth.vm.logic.arithmetic

3. Added all test cases from the EIP

4. Added opcodes that expose the functionality to the constantinople opcodes

Cute Animal Picture

[pipermerriam]

Big 👍 for adding individual opcode function tests.

[cburgdorf]

@pipermerriam failing CI seems like something related to the new network abstraction and unrelated to this PR as far as I can tell. https://circleci.com/gh/ethereum/py-evm/50905?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-build-link

[pipermerriam]

@cburgdorf see cburgdorf#1 for how to do this without the library.

[pipermerriam]

Regarding test failure: ethereum/pytest-asyncio-network-simulator#6 incoming

[pipermerriam]

@cburgdorf you should be able to rebase this now and pass CI

[cburgdorf]

@pipermerriam yep, passes! There's one more shifting opcode left to implement and I also want to put more work in the tests. I'll try to finish that up tomorrow.

[carver]

Can we start running selected Constantinople state tests now?

[pipermerriam]

Can we start running selected Constantinople state tests now?

@carver yes, but my recommendation would be to do that in a separate PR. We'll need to update the commit that our fixtures are current set at (probably to whatever is in geth master?) and then fix anything that is broken.

[cburgdorf]

@pipermerriam @carver Alright, I decided to branch of #1134 for the SAR opcode because I really need to ramp up my unicon bits & bytes skills for that one (Or maybe I just need to pick your brain more @carver 😅)

So, this one is ready for review if anyone is up for it.