(WIP) Call fixes precompiles

[mratsim]

Fix #122, required by #121

I'm opening it the PR to start discussion implementation (or things not implemented) and tests.

---

A change in eth_trie file structure broke the CI: https://travis-ci.org/status-im/nimbus/jobs/437522149#L1050

[zah]

If this proc is generic now (due to the static param), it would make sense to extract the parts of it that don't depend on the opCode param in helper procs (the goal would be the reduce the overall code size).

[tersec]

When the GeneralStateTests.md changes show up, it would also be good to add the newly passing tests to test_generalstate_failing.nim to prevent regressions.

[tersec]

Where do the PrecompileTests come from? I don't see them in https://github.com/ethereum/tests

[tersec]

Do you have any insight as to why these PrecompileTests mostly work, but the state tests in stPreCompiledContracts2 don't?

[coffeepots]

@tersec Calling for review for this work to be merged before continuing on the state tests.
In response to your query, I'm not sure why the stPreCompiledContracts2 don't pass yet but it's worth noting there is currently an issue in the modexp precompile that needs fixing. However this branch still pushes forward more passing tests so I believe it's worth doing these fixes in another PR to prevent more merge conflicts going forward.

[mratsim]

Regarding modExp are some tests passing and other failings or are all failing?

[coffeepots]

@mratsim ModExp currently fails all tests except one. Current issues are:

1. Consuming Data

Fails assertion when consuming data here: https://github.com/status-im/nimbus/blob/0366dd47d94c07e7d32606343adfda6043b7a59d/nimbus/vm/precompiles.nim#L117

base = rawMsg.rangeToPaddedUint256(96, start_exp - 1) fails an assertion check in io.nim because some tests use a base_len that is >32.

base_len values in the tests are 32, 64, 128, 256, 512 and 1024.

2. Results

The remaining tests that don't have the above error are eip_example1 and eip_example2

eip_example1:
Output : 0000000000000000000000000000000000000000000000000000000000000000
Expected: 0000000000000000000000000000000000000000000000000000000000000001

eip_example2: Is the only test that passes, but is expecting zero output anyway.

[tersec]

@tersec Calling for review for this work to be merged before continuing on the state tests.
In response to your query, I'm not sure why the stPreCompiledContracts2 don't pass yet but it's worth noting there is currently an issue in the modexp precompile that needs fixing. However this branch still pushes forward more passing tests so I believe it's worth doing these fixes in another PR to prevent more merge conflicts going forward.

I agree.