The 'chain' and 'pins' rule sets diverge on exactly one shape: a frame that carries value with no target. No captured transaction has that shape, so the positive branch of the divergence survey runs only against constructed input, never against bytes the chain produced.
This is the coverage gap with the most to say, because it is the one case where the deployed binary and the pinned EIP text charge differently, and a real transaction would settle which figure the nodes apply.
What to do: watch for such a transaction, capture it with bunx tsx scripts/capture-fixtures.ts [from-block] [to-block], and add the file under test/fixtures/chain/. Report the result either way: if compareRuleSets on the captured transaction disagrees with what the node charged, that is a divergence report, not a test to adjust.
Context is in docs/OPEN-ITEMS.md. CONTRIBUTING.md covers the oracles and why captured data is treated as the authority.
The
'chain'and'pins'rule sets diverge on exactly one shape: a frame that carries value with no target. No captured transaction has that shape, so the positive branch of the divergence survey runs only against constructed input, never against bytes the chain produced.This is the coverage gap with the most to say, because it is the one case where the deployed binary and the pinned EIP text charge differently, and a real transaction would settle which figure the nodes apply.
What to do: watch for such a transaction, capture it with
bunx tsx scripts/capture-fixtures.ts [from-block] [to-block], and add the file undertest/fixtures/chain/. Report the result either way: ifcompareRuleSetson the captured transaction disagrees with what the node charged, that is a divergence report, not a test to adjust.Context is in
docs/OPEN-ITEMS.md.CONTRIBUTING.mdcovers the oracles and why captured data is treated as the authority.