EIP-8037 Alignment for bal-devnet-7

Latest EIP changes made at the end of Soldøgn interop have not yet been reflected on our EIP-8037 implementation. Apart from this, there are a few bugs / issues caught in bal-devnet-6 that need to be addressed.

Latest EIP-8037 Spec Changes

• Update EIP-8037: fixed CPSB + frame accounting EIPs#11573
• Update EIP-8037: Update parameters and rationale section EIPs#11616
• Update EIP-8037: Fix bugs from bal-devnet-6 EIPs#11611

Test Discrepancies Reported

For transparency, and to aid in aligning with the latest spec, these bugs were highlighted by @rakita/@benaadams & ethrex. PRs to address the spec/test updates below:

• Fix halt sematics to refund state gas to top call correctly
  • feat(spec-specs, tests): EIP-8037 halt refunds spilled state gas #2815
• Tx create on halt/revert refund is not done @spencer-tb
  • feat(spec-specs, tests): EIP-8037 more refund fixes #2823
• block.state_gas_used does not include eip7702 state refund
  • fix(spec,test): 7702 refund block-level gas accounting for EIP-8037 #2816 @fselmo
• State refund propagated by mistake if sstore 0 -> x -> 0, revert is done in one frame @spencer-tb
  • feat(spec-specs, tests): EIP-8037 more refund fixes #2823
• Tx create collision ported static test (not possible to happen on chain) regular gas was not consumed
  • Added a specific 8037 test for this feat(spec-specs, tests): EIP-8037 more refund fixes #2823
• Update parameters: update the state byte cost params and update the CPSB (1174 -> 1530)
  • feat(spec,test): update bal-devnet-7 EIP-8037 bytes values; CPSB 1174 -> 1530 #2827

In addition to these, perhaps worth noting from @benaadams:

• Create + Selfdestruct initcode; but I think test being done for that by ethrex? (edit: This scenario may not be the case, as reported by Ivan from Ethrex they are no longer working on a test here -- needs investigation)
  • feat(spec-specs, tests): EIP-8037 tx created contracts destroyed in same tx #2828