Update to ethereumjs-common v1.1.0

[danjm]

This PR updates code and package.json to use ethereumjs-common v1.1.0

[coveralls]

Coverage remained the same at 73.818% when pulling 9dbb00b on update-ethereumjs-common-version into 69c73b1 on master.

[whymarrh]

LGTM

[holgerd77]

I think it would make sense to directly update this to v1.1.0 (so ^1.1.0 (once released), then the library will be more safe to be used in a Petersburg context and we can directly state this in the subsequent release notes here.

@danjm If you want you can directly update to safe time. I can then re-trigger CI once the new Common version is released.

[danjm]

@holgerd77 Updated to ^1.1.0

[holgerd77]

Ok, various difficulty related tests failing, maybe that's an issue introduced in the hardfork lineup in Common along with the Petersburg HF, will also investigate a bit. Eventually we have to do a follow-up bugfix release (I would suspect some HF comparison logic bug?).

[holgerd77]

Some in-between output, here I check for the hardfork chosen in the BlockHeader.canonicalDifficulty() function as well as how the greatThanEqual Common function evaluates:

Chosen hardfork: constantinople
> Constantinople?: true
> Byzantium?: true
not ok 23313 test canonicalDifficulty (fork determination by block number)
  ---
    operator: equal
    expected: '4841178247367196615'
    actual:   '4841178247367192519'
    at: runDifficultyTests (/Users/hdrewes/Documents/DEV/EthereumJS/ethereumjs-block/tests/difficulty.js:19:7)
    stack: |-
      Error: test canonicalDifficulty (fork determination by block number)
          at Test.assert [as _assert] (/Users/hdrewes/Documents/DEV/EthereumJS/ethereumjs-block/node_modules/tape/lib/test.js:224:54)
          at Test.bound [as _assert] (/Users/hdrewes/Documents/DEV/EthereumJS/ethereumjs-block/node_modules/tape/lib/test.js:76:32)
          at Test.equal (/Users/hdrewes/Documents/DEV/EthereumJS/ethereumjs-block/node_modules/tape/lib/test.js:384:10)
          at Test.bound [as equal] (/Users/hdrewes/Documents/DEV/EthereumJS/ethereumjs-block/node_modules/tape/lib/test.js:76:32)
          at runDifficultyTests (/Users/hdrewes/Documents/DEV/EthereumJS/ethereumjs-block/tests/difficulty.js:19:7)
          at Test.t (/Users/hdrewes/Documents/DEV/EthereumJS/ethereumjs-block/tests/difficulty.js:67:7)
          at Test.bound [as _cb] (/Users/hdrewes/Documents/DEV/EthereumJS/ethereumjs-block/node_modules/tape/lib/test.js:76:32)
          at Test.run (/Users/hdrewes/Documents/DEV/EthereumJS/ethereumjs-block/node_modules/tape/lib/test.js:95:10)
          at Test.bound [as run] (/Users/hdrewes/Documents/DEV/EthereumJS/ethereumjs-block/node_modules/tape/lib/test.js:76:32)
          at Immediate.next (/Users/hdrewes/Documents/DEV/EthereumJS/ethereumjs-block/node_modules/tape/lib/results.js:71:15)
  ...

This looks ok, nevertheless the difficulty if falsely calculated. I suspect the params read from Common subsequently.

[holgerd77]

Ok, got it: this is failing because we run the difficulty tests on Ropsten in difficulty.js on difficultyRopstenByzantium.json:

const chainTestData = {
    'mainnet': require('./difficultyMainNetwork.json').tests,
    'ropsten': require('./difficultyRopstenByzantium.json').tests
  }

Since the new Common release also defines a hardfork number for Constantinople for Ropsten and block numbers seems to be above the threshold, difficulty is calculated within the Constantinople context.

If test file is changed to difficultyRopstenConstantinople.json (file already included in the tests folder), tests pass again.

Can you update this here?

[danjm]

@holgerd77 Thanks for the investigation on the test failures. I've made the changes and tests are now passing.

[holgerd77]

Ok, will directly merge and prepare a release.

[holgerd77]

Looks good.