Process same-slot slashings before builder payments in Gloas #4726
Conversation
0xc0de42
force-pushed
the
fix_same_slot_slashing
branch
4 times, most recently
from
37d0a21 to
e0d251b
Compare
0xc0de42
force-pushed
the
fix_same_slot_slashing
branch
from
e0d251b to
fe3e8d7
Compare
0xc0de42
force-pushed
the
fix_same_slot_slashing
branch
from
fe3e8d7 to
68a9c02
Compare
|
|
||
| @with_gloas_and_later | ||
| @spec_state_test | ||
| def test_process_block_then_slash_non_timely_builder(spec, state): |
There was a problem hiding this comment.
I think it would be good if you yield so it would create a test vector
There was a problem hiding this comment.
Perphas one of the operations formats here: tests/formats/operations/README.md
jtraglia
changed the title
|
Hey @0xc0de42, I'm confused. This PR doesn't do this:
|
|
Hey @jtraglia , thanks for the response. The PR doesn't do that, I mentioned this topic as it was suggested to do so at issue #4561. Sorry for the confusion.
I was not sure whether specs should cover both cases:
This test case only addressed the second case. |
| yield "block", block | ||
| spec.process_block(state, block) | ||
| # Slash validator | ||
| spec.slash_validator(state, builder_index) |
There was a problem hiding this comment.
This state change would have to be after the post, but I think it is better to use another format that sorrounds the slashing
3 participants
Fix processing of same-slot slashings before builder payments in Gloas (#4561)
Related to #4561
Slashing currently only succeeds when the validator is slashed after the block has been processed.
This behavior fails earlier because of the assert not builder.slashed check in process_execution_payload_bid(state, block).
Reordering the internal function calls within process_block. For example, evaluating process_block_header at the end, does not resolve the issue, as the slashing condition is still enforced before the builder payment is applied.