Skip to content

EIP-7708: ETH transfer logs - #2873

Open
anvacaru wants to merge 13 commits into
masterfrom
eip-7708
Open

EIP-7708: ETH transfer logs#2873
anvacaru wants to merge 13 commits into
masterfrom
eip-7708

Conversation

@anvacaru

Copy link
Copy Markdown
Contributor

Implements EIP-7708: ETH balance changes emit synthetic logs (CALL-value transfers, SELFDESTRUCT sweeps, fee payments).

Changes:

  • evm.md: transfer-log emission in #transferFunds; schedule.md: Ghaseip7708. (The Gtransferlog intrinsic pricing will be implemented in a follow-up PR )

anvacaru and others added 13 commits July 24, 2026 15:46
….py: parse Amsterdam block header (blockAccessListHash, slotNumber); discard receipts/blockAccessList fixture keys in the harness
… max 21 / fraction 11684671)

AMSTERDAM inherited the Osaka/Prague blob parameters (target 6 / max 9 /
fraction 5007716). EELS amsterdam vm/gas.py has BLOB_SCHEDULE_TARGET = 14
and BLOB_SCHEDULE_MAX = 21 (at PER_BLOB = 2^17: Gtargetblobgas = 1835008,
Gmaxblobgas = 2752512) and BLOB_BASE_FEE_UPDATE_FRACTION = 11684671,
matching the mainnet BPO2 blob schedule (EIP-8135 / EIP-7892) that
Amsterdam starts from.

Raising Gmaxblobgas from 9 to 21 blobs exposes the pre-existing missing
EIP-7594 per-tx blob limit (EELS BLOB_COUNT_LIMIT = 6, fork.py:150):
the blob_count_21 sub-case of invalid_max_blobs_per_tx was previously
rejected by the too-small inherited block max instead of the per-tx
check. Mask it in failing.llvm next to its already-masked blob_count_7/8
siblings until the per-tx limit is implemented.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
read_csv_file crashed test collection with IndexError on any blank
line in failing.llvm/slow.llvm — which the fixtures-failing-list
Makefile target itself can produce via its sort/sed post-processing.
Skip empty rows.
# Conflicts:
#	kevm-pyk/src/kevm_pyk/kproj/evm-semantics/driver.md
Emit an EIP-7708 transfer LOG3 at every nonzero-value transfer to a different
account. All four categories (tx-level, CALL, CREATE endowment, SELFDESTRUCT)
route through #transferFunds, so the log is appended in its single balance-
moving rule (ACCTFROM =/= ACCTTO), gated on the new Ghaseip7708 flag (true at
AMSTERDAM) and VALUE =/= 0. The entry matches the EELS reference exactly:
address = SYSTEM_ADDRESS (0xff..fe), topics = (keccak256
"Transfer(address,address,uint256)", from, to), data = 32-byte big-endian
value. No new gas charge (EIP-2780 already prices Gtransferlog). Reverts are
handled by the existing #pushWorldState/#popWorldState substate snapshot (the
snapshot precedes the transfer), so failed CALL/CREATE frames drop the log.

Unmask 34 of the 37 eip7708 fixtures. The remaining amsterdam fixtures are
blocked by pre-existing non-7708 gaps (CREATE state-gas #8, delegation/typed-tx
Ctxfloor #6) and the transition fixture by the single-schedule harness limit;
verified the transfer logs themselves are correct in those cases. osaka/prague
sweep unchanged (5 fail / 5122 pass); amsterdam-native suites show only the 10
known pre-existing gap failures.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Base automatically changed from amsterdam-foundation to master September 8, 2026 11:52
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant