consensus, core, tests: implement Metropolis EIP 649

[karalabe]

Implements ethereum/EIPs#669.

The PR also pulls in the latest test suite, which among other throws out explicit log validation to implicit one via RLP(logs).

---

Note, some tests are failing because the consensus test suite was only partially updated to the new log format, so the non-updated ones cannot be parsed. I'll keep checking the test suite to see if there are updates to fix them.

[fjl]

Please remove the logs check so we can merge.

[karalabe]

I'll wait to see how many fail (just updates the test suite), and will disable only the failing tests (so we're not overly zealous and accidentally miss consensus issues).

[holiman]

This was already here, not new for this PR: https://github.com/ethereum/go-ethereum/pull/15028/files#diff-e226000964244889df127e5b44dbf192R546 :

 		r.Div(blockReward, big32)
 		reward.Add(reward, r)

Looks un-intuitive (reusing r which has a totally different meaning on the lines before) and unnecessary (always divide blockReward by 32).
I'd prefer:

	blockReward := frontierBlockReward
        newphewReward := frontierNewphewReward
 	if config.IsMetropolis(header.Number) {
 		blockReward = metropolisBlockReward
                nephewReward:= metropolisNephewReward
 	}
       ...
 		reward.Add(reward, nephewReward)

[devsnd]

@karalabe I noticed that the big299999 is 299_000 and not 3_000_000 (-1) as in the EIP

https://github.com/ethereum/go-ethereum/pull/15028/files#diff-e226000964244889df127e5b44dbf192R310
https://github.com/ethereum/go-ethereum/pull/15028/files#diff-e226000964244889df127e5b44dbf192R351

Is this on purpose?

[holiman]

@devsnd OMG wow good catch!

It's using 299999 instead of 2999999

[karalabe]

@fjl I've disabled the 3 failing test files, we can readd them later.
@holiman I've fixed the 2_999_999 issue, PLTA.
@devsnd 2,999,999 is correct because the code there is working on the parent block number, not the actual block number. Though do ping me if you still think it's wrong.

[holiman]

Just to be clear, my initial comment above was to point out that it was supposed to be one less than 3M, but then I understood that @devsnd had correctly spotted an error - so again, good catch!

[holiman]

PTAL:ed, looks good to me