[AgentME]

If it's true that Ethereum blocks have any kind of checksum of the expected results, then if the patch has a mistake, I'd think the failure mode would just be that they continue to fail processing the failed block, or they get further but fail to process a more recent block. In that case, then I think it would make sense here to rush out a patch that's been manually checked, and then go back to add some tests.

[xiphias2]

The right way to fix this is to have an integration test that passes on all other implementations but fails on OpenEthereum, and push this fix afterwards.

People are storing and trading billions of dollars on Ethereum, so I think both automated and manual checking is important. Also an integration test for all branches should be able to find these errors.