tests: Support EIP-7892 - BPO forks

[pdobacz]

https://eips.ethereum.org/EIPS/eip-7892

Closes #1325 and follows the proposal there.

Blob parameter only forks. This ended up a considerable revamp, as we've moved on from hardcoded blob schedules to ones coming from JSON config (post Osaka). Because of this, a bunch of internal APIs needed to be shuffled, I hope this all still makes sense.

To ease review, the "main" commit follows the old naming convention, which seems to have bin a misnomer - it took a single entry of a blob schedule for the blob schedule. The two follow-up commits, to be squashed away before merging, clean this up.

But, this PR should also be reviewable as a whole, just need to follow the new naming.

Filling with evmone-t8n requires an update to EEST: ethereum/execution-spec-tests#2264

[pdobacz]

NOTE: this used to be constexpr but it cannot be anymore, since I've decided to overload get_blob_params. I don't think the constexprness has been leveraged anywhere, and now it also throws so probably doesn't even make sense to be constexpr.. But yea, lmk if this a problem

[codecov]

Codecov Report

❌ Patch coverage is 81.60920% with 32 lines in your changes missing coverage. Please review.
✅ Project coverage is 87.24%. Comparing base (008669c) to head (918b013).
⚠️ Report is 1 commits behind head on master.

Files with missing lines	Patch %	Lines
test/unittests/state_block_test.cpp	16.66%	0 Missing and 15 partials ⚠️
test/utils/utils.cpp	0.00%	4 Missing and 4 partials ⚠️
test/statetest/statetest_loader.cpp	89.74%	2 Missing and 2 partials ⚠️
test/state/blob_schedule.cpp	90.32%	2 Missing and 1 partial ⚠️
test/blockchaintest/blockchaintest_runner.cpp	90.00%	0 Missing and 1 partial ⚠️
test/t8n/t8n.cpp	94.11%	0 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1330      +/-   ##
==========================================
- Coverage   87.25%   87.24%   -0.01%     
==========================================
  Files         169      169              
  Lines       24864    24970     +106     
  Branches     4081     4106      +25     
==========================================
+ Hits        21695    21786      +91     
- Misses        521      529       +8     
- Partials     2648     2655       +7     

Flag	Coverage Δ
eest-develop	79.87% <87.12%> (+0.05%)	⬆️
eest-develop-gmp	15.97% <33.12%> (+0.13%)	⬆️
eest-legacy	11.22% <36.20%> (+0.11%)	⬆️
eest-legacy-silkpre	17.63% <28.02%> (+0.07%)	⬆️
evmone-unittests	83.81% <64.94%> (-0.08%)	⬇️

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

Components	Coverage Δ
core	93.58% <91.42%> (-0.02%)	⬇️
tooling	88.08% <84.61%> (-0.22%)	⬇️
tests	84.14% <68.75%> (+0.01%)	⬆️

Files with missing lines	Coverage Δ
test/blockchaintest/blockchaintest_loader.cpp	95.83% <100.00%> (+0.21%)	⬆️
test/state/block.cpp	100.00% <100.00%> (ø)
test/state/block.hpp	100.00% <ø> (ø)
test/statetest/statetest.hpp	85.71% <ø> (ø)
test/statetest/statetest_runner.cpp	80.43% <100.00%> (+0.43%)	⬆️
test/unittests/state_transition.cpp	81.30% <100.00%> (ø)
test/unittests/state_transition_block_test.cpp	100.00% <100.00%> (ø)
test/unittests/state_transition_call_test.cpp	100.00% <100.00%> (ø)
test/unittests/state_tx_test.cpp	83.76% <100.00%> (+0.17%)	⬆️
...est/unittests/statetest_loader_block_info_test.cpp	72.88% <100.00%> (+1.09%)	⬆️
... and 6 more

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[pdobacz]

I'm puzzled by why do things like unittests or t8n still come up on the coverage report, I left them on the ignore list.

This is the last thing to tackle here I guess, but otherwise ready for review

EDIT: of course, the ignore is for the EEST job only, that's why t8n comes up. I also filled some coverage gaps. I think the report is fine, there are some immaterial or irrelevant lines missed, some of which we will work on filling later.

[chfast]

I think the general direction is good: don't hardcode the blob schedule in EVM.

[chfast]

Focusing on the interface for EVM here, I think the mapping should be rather timestamp => BlobParams so this is independent of the revision and the network name.

So we should use something like flat_map to represent BlobSchedule.

Getting the BlobSchedule by name or from JSON should be the job for the EVM user. I imagine this will be hardcoded for Mainnet and constructed from JSON for tests.

I'm not sure how much refactoring this is, so we can leave it for some other task. Maybe when we will be putting better boundary between test and consensus code.

[chfast]

Now I realized that passing BlobParams instead of BlobSchedule is more in-line with passing revision.

[pdobacz]

Focusing on the interface for EVM here, I think the mapping should be rather timestamp => BlobParams so this is independent of the revision and the network name.

So I thought about this today. What troubles me, is whether or not we'd still want to fetch BlobParams by fork explicitly, bypassing timestamp. More generally, I'm not 100% sure about the boundary between EEST and evmone here - in the contexts of state/blockchaintests, t8n, and potentially the future Engine API test consumption.

I think it's ok to merge as is to tick off coverage work, but I can try the refactor afterwards and we can decide the API considerations with the testing team.

Now I realized that passing BlobParams instead of BlobSchedule is more in-line with passing revision.

Not sure what do you mean? passing where?

[chfast]

I meant the same as your concern.

[chfast]

You can rework commits and we can merge this.

[pdobacz]

@chfast I've tacked on a commit implementing the change we've discussed (using resolved blobParams), PTAL, and if it's OK I'll squash again and we'll be ready to merge.

[chfast]

Make sure the final commit message will be up to date.