CREATE2 implementation in LegacyVM #5121
Conversation
Codecov Report
@@ Coverage Diff @@
## develop #5121 +/- ##
===========================================
+ Coverage 58.39% 58.47% +0.08%
===========================================
Files 336 337 +1
Lines 26973 27017 +44
Branches 3161 3162 +1
===========================================
+ Hits 15750 15798 +48
+ Misses 10154 10152 -2
+ Partials 1069 1067 -2 |
|
Somehow the test succeeds on Windows but generates |
|
Oh wow, switch statement in VM is implemented with some gcc extention magic on *nixes, I didn't notice it |
|
Still some details are not finalized:
Ortherwise this can be merged I guess. |
| @@ -365,7 +365,7 @@ bool Executive::createOpcode(Address const& _sender, u256 const& _endowment, u25 | |||
|
|
|||
| bool Executive::create2Opcode(Address const& _sender, u256 const& _endowment, u256 const& _gasPrice, u256 const& _gas, bytesConstRef _init, Address const& _origin, u256 const& _salt) | |||
| { | |||
| m_newAddress = right160(sha3(_sender.asBytes() + toBigEndian(_salt) + sha3(_init).asBytes())); | |||
| m_newAddress = right160(sha3(_sender.asBytes() + toBigEndian(_salt) + _init)); | |||
There was a problem hiding this comment.
I think the consensus is going towards adding the 0xff prefix. The benefit is that old addresses can't clash with new ones, which is a win because it prevents a class of corner cases that might appear in old create code.
There was a problem hiding this comment.
I hope we will make the final decision in the next Add Dev Call. Unless there is a way to make it sooner.
#5119