Replay for Byzantium data #436
Conversation
| if transfersValue && !evm.chainRules.IsEIP4762 { | ||
| if transfersValue && (!evm.chainRules.IsEIP4762 || replayMode) { |
There was a problem hiding this comment.
Other case that we should disable the effects.
There was a problem hiding this comment.
but isn't just deactivating eip 4762 enough?
There was a problem hiding this comment.
There was a problem hiding this comment.
note to self: no it isn't, because if you do then eip2929 is still active and we don't want to handle that case. I guess eip2929 becomes the new final boss.
| bc.stateCache.InitTransitionStatus(true, true) | ||
| bc.stateCache.EndVerkleTransition() | ||
| bc.stateCache.InitTransitionStatus(false, false) | ||
| // bc.stateCache.EndVerkleTransition() |
There was a problem hiding this comment.
This is a required change to avoid an error when processing the second block.
Before merging we have to figure out a coherent way to do this for both replay and non-replay setups to work correctly.
There was a problem hiding this comment.
indeed. but if it's the only change that we have to keep in a separate branch, that's ok too.
| for i := prevNumber; i > 0 && i >= prevNumber-params.Eip2935BlockHashHistorySize; i-- { | ||
| for i := prevNumber; i > 0 && i > prevNumber-params.Eip2935BlockHashHistorySize; i-- { |
There was a problem hiding this comment.
Can be removed when #438 is merged and the base branch change bubbles up when we rebase branches. For now leaving the fix since it's needed.
core/vm/interpreter.go
Outdated
There was a problem hiding this comment.
note to self: these changes are why this can't be merged, it's only to replay very old stuff.
jsign
force-pushed
the
jsign-replay-kaustinen6
branch
from
6d11543
to
e796a78
Compare
Signed-off-by: Ignacio Hagopian <jsign.uy@gmail.com>
Signed-off-by: Ignacio Hagopian <jsign.uy@gmail.com>
Signed-off-by: Ignacio Hagopian <jsign.uy@gmail.com>
Signed-off-by: Ignacio Hagopian <jsign.uy@gmail.com>
Signed-off-by: Ignacio Hagopian <jsign.uy@gmail.com>
jsign
force-pushed
the
jsign-replay-kaustinen6-for-old
branch
from
b524731
to
cd92db3
Compare
Signed-off-by: Ignacio Hagopian <jsign.uy@gmail.com>
This PR is basically the branch from #431 but with further adjustments to allow trying to replay the old (smaller) replay data.
This PR isn't intended to be merged, but was to be able to fix anything that was a bug for replay to work in general. #431 configuration reg precompiles/instructions shouldn't be changed. At some point we need to decide to merge bug-related/adjustements to that branch, see comments.