core/vm: implement eof opcodes #30511
Conversation
MariusVanDerWijden
requested review from
fjl,
holiman,
karalabe,
gballet,
rjl493456442,
zsfelfoldi,
s1na and
lightclient
as code owners
|
Merged eof-validation, rebase time :) |
MariusVanDerWijden
force-pushed
the
eof-opcodes
branch
from
f6dd3c8 to
73a3a54
Compare
|
Rebased, looks like I broke some verkle stuff, will fix later |
Move EOF activation to Osaka
Update the EOF container parsing to allow the PC to use the same semantics as the legacy EVM. Also, a new test container maker is necessary to handle the particulars of the unit tests, mostly unrelated to
Move EOF interpreter to use same PC semantics as legacy EVM
|
|
||
| // Validate initcode per EOF rules. If caller is EOF and initcode is legacy, fail. | ||
| isInitcodeEOF := hasEOFMagic(codeAndHash.code) | ||
| if isInitcodeEOF { |
There was a problem hiding this comment.
We could think about moving this a bit later. Right now we do the heavy things (unmarshalling and validation) before we actually deduct the gas and check for nonce overflow
|
Hey folks, I’m an R&D engineer at Tenderly, and I’ve been following the work on this PR for some time now. We’ve been working on integrating it locally, so we’d really appreciate your input. |
|
Hey @veniger, I just came back from holiday. I would like to merge this soon, probably after we updated it to devnet-1 spec and the pectra fork has passed |
|
🫡 |
based on #30418
rework of #29518
I could not test this properly yet, as the validity of the blockchain tests depend on the prague spec. In order to test this correctly, we need to upstream 6110 and 7002
Unfortunately I couldn't preserve the authorship, I think by using the coauthored tag, people will receive their due credit once this is merged.
Co-authored-by: lightclient [email protected]
Co-authored-by: shemnon [email protected]
Co-authored-by: holiman [email protected]