Closed pdobacz closed 2 months ago
It depends who is complaining about the missing accessLists
. If only evmone, then we can just relax the parsing rules. All this is likely because we are forced to guess the transaction type by the fields...
All modified and coverable lines are covered by tests :white_check_mark:
Project coverage is 94.21%. Comparing base (
c885db3
) to head (86fc519
). Report is 1 commits behind head on master.
It depends who is complaining about the missing
accessLists
. If only evmone, then we can just relax the parsing rules. All this is likely because we are forced to guess the transaction type by the fields...
Nope, it's consumers of the exported tests.
FWIW I diffed the 0.13.0 tests and ones exported from this PR branch and the only difference is the addition of some empty accessLists
fields in some JSONs, which seemed to already contain 1559 fields
Ok, let's go with this. The only alternative I have is to do the tx type guessing during export...
Just add a comment why this is the default transaction type.
Opening as draft to discuss if I'm understanding things correctly here.
The problem was that our exported test txs were missing
accessLists
field, even though they containedeip1559
fields. I think making the default state_transition tx be eip1559, and making it explicit wherever we want legacy instead is the solution, but I'm not sure.Example exported JSON without
accessLists
:from
fixtures/state_tests/state_transition/eof_examples_callf.json
.Does this make sense?