Update ethereum-tests submodule to 8.0.0 / Remove EIP-2315 from berlin

[holgerd77]

This is a simple PR which updates the ethereum-tests submodule to the Berlin release 8.0.0 with a removed EIP-2315 HF. EIP-2315 is also removed here in Common in the berlin HF definition file.

I originally also wanted to add the block numbers for berlin along this PR. But I don't trust these numbers yet since the Ropsten HF would be so close and there is no corresponding go-ethereum release out yet, see also ethereum/pm#248 (comment).

[holgerd77]

Note that this PR is also running with the type: test all hardforks label.

[codecov]

Codecov Report

Merging #1142 (ffaf9cb) into master (94651f7) will increase coverage by 0.07%.
The diff coverage is 100.00%.

Flag	Coverage Δ
block	81.82% <ø> (+0.21%)	⬆️
blockchain	84.19% <ø> (+0.06%)	⬆️
client	87.04% <ø> (+0.19%)	⬆️
common	86.47% <ø> (-1.41%)	⬇️
devp2p	84.28% <ø> (ø)
ethash	82.08% <ø> (ø)
tx	84.86% <ø> (+0.18%)	⬆️
vm	81.67% <100.00%> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

[holgerd77]

Some tests failing, feel very free to directly push here. Will stop for now.

[jochem-brouwer]

Have given this a quick glance, this is probably the problem. But I don't have time today to actually check it.

[jochem-brouwer]

This should fix the problem. Have also moved EIP2315 tests from berlin directory into EIPs directory.

@holgerd77 as for why your tests passed locally: you probably forgot to rebuild Common, so if you would create a Berlin common it automatically activates EIP2315.

[holgerd77]

@jochem-brouwer lol, good point 😄 (yes, you're right)

[holgerd77]

@jochem-brouwer we can now also include the Berlin HF blocks here along since this is now settled. Do you might want to directly do, don't want to push here in between?

[jochem-brouwer]

Ah right, good point, will add them here and push.

[holgerd77]

@jochem-brouwer cool! 👍

Have also cross-checked the EIP-2315 implementation, I think we've all covered now.

[jochem-brouwer]

Right, I don't have the fork hashes, and sadly we also don't have an utility module to generate these fork hashes (do we want this?). I have asked on R&D if someone has the hashes for me.

[holgerd77]

@jochem-brouwer ah, you are not reading my messages 😋 - also wrote about this in the Yolov3 PR - you can use the Common._calcForkHash() method for that.

[holgerd77]

(so put in the block numbers and then run with a Common set to berlin from console)

[jochem-brouwer]

Ah I missed this message, just figured this out as well that we actually have this feature 😅

[jochem-brouwer]

Found a possible problem: we do not pad the fork hashes, which is why for instance this kovan hash is of 7 characters (9 characters with 0x prefixed). This is problematic if we do Buffer.from(hash.slice(2)), you'd get the Buffer 10ffe5, where we expect Buffer 010ffe56.

I don't know if this is an issue, but it is not consistent with other packages, we usually left pad these strings or Buffers.

Anyways, have added Berlin blocks to Common, see the EF blog for block numbers. Also got the fork hashes from Geth's fork hash tests. And I also learned we can actually calculate these hashes 😅

[holgerd77]

Ok, double-checked the numbers, LGTM