ci: disabling most of the hardfork-specific tests

[evertonfraga]

As promised, this PR disables most of the hardfork specific tests, to favor job concurrency.

It can be easily reverted, by un-commenting those lines.

[codecov]

Codecov Report

Merging #950 (5d8639d) into master (d8e7066) will decrease coverage by 0.02%.
The diff coverage is n/a.

Flag	Coverage Δ
block	75.04% <ø> (ø)
blockchain	77.39% <ø> (ø)
common	92.11% <ø> (ø)
ethash	82.08% <ø> (ø)
tx	86.04% <ø> (-0.22%)	⬇️
vm	87.21% <ø> (ø)

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

[evertonfraga]

@jochem-brouwer double-checking: what are the most important HFs to test here? We can keep to 2 or 3 for brevity.

@holgerd77 After choosing the HFs, can I tweak the required status checks for master branch?

[jochem-brouwer]

I'd say the most important are MuirGlacier (mainnet), Istanbul (latest hard fork w/h difficulty EIP -> MuirGlacier) since most networks run on one of both. All other forks are "outdated" so they don't need to be stress tested on every PR (except if changes are specific for that HF).

[jochem-brouwer]

I'd say we also include MuirGlacier since this is what currently runs on mainnet.

[evertonfraga]

Thanks! the changes are in place.

[jochem-brouwer]

Yup, great!

[evertonfraga]

I have updated the required checks according to the state and blockchain I have just changed.

Closing and reopening PR to try to see the Required Checks change in effect.

[holgerd77]

All other forks are "outdated" so they don't need to be stress tested on every PR (except if changes are specific for that HF).

I would be careful with this, there are plenty of ways to break old HF code with lastest-fork-feature-adding PRs, we just had an example for this with the broken Constantinople tests fixed in #931. Should be ok though I would say with #951, we should nevertheless be aware of this and in doubt rather trigger the test runs for all the hardforks.