feat: implement the v10 migration #142
Conversation
arajasek
force-pushed
the
asr/migration
branch
from
140cda3 to
0991928
Compare
arajasek
force-pushed
the
asr/migration
branch
from
0991928 to
08baf70
Compare
| return cid.Undef, xerrors.Errorf("failed to make eth zero address: %w", err) | ||
| } | ||
|
|
||
| migrations[oldInitCodeCID] = initActorMigrator{OutCodeCID: newInitCodeCID, EthZeroAddress: ethZeroAddr} |
There was a problem hiding this comment.
Oh, I see. Yeah.... that's not so fun. Could we just apply it as an implicit message? Or is that not allowed.
There was a problem hiding this comment.
What's not so fun? Manipulating the init actor state like this? It's...not too different from stuff what we've done before.
We could apply an implicit message after the migration has run (so in Lotus, after we've done all the code CID transformations and such). But I'm generally comfortable with this (honestly we did much more in the v17 migration).
| return cid.Undef, xerrors.Errorf("failed to set EAM Actor: %w", err) | ||
| } | ||
|
|
||
| // Create the EthZeroAddress as an EthAccount |
There was a problem hiding this comment.
I'm not seeing us create it. Are we just checking to see if it exists?
There was a problem hiding this comment.
WDYM? There's basically 2 steps to "creating" it:
- Assigning it an ID in the init actor (happens as part of the init actor migration)
- Adding that ID to the state tree as an EthAccount, which happens ~25 lines after this comment.
We're not assigning it a robust address, but that seemed okay to me?
There was a problem hiding this comment.
Assigning it an ID in the init actor (happens as part of the init actor migration)
Ah, this is what I was looking for. You're right, we're creating the actor here, we've already assigned the address.
Addressfield nil