Tests for EIP-3855: PUSH0 instruction #1000
Conversation
gumb0
force-pushed
the
push0
branch
2 times, most recently
from
0d19c77 to
cb76344
Compare
|
|
||
| 0000000000000000000000000000000000000100: | ||
| balance: 0 | ||
| code: ':raw 0x5f' |
There was a problem hiding this comment.
save gas before, do (PUSH0) in lllc https://github.com/winsvega/solidity (add as a PR here, retesteth docker uses it)
save the gas after
sstore the gas used change into storage.
| b94f5374fce5edbc8e2a8697c15331677e6ebf0b: | ||
| storage: | ||
| 0x01: 0x01 | ||
| # Same expectation, but for London execution fails on 5f opcode, for 3855 it fails with stack overflow |
There was a problem hiding this comment.
new opcode to be added into opcode tests by Ori.
https://github.com/ethereum/tests/tree/develop/src/GeneralStateTestsFiller/stBadOpcode
and here https://github.com/ethereum/tests/tree/develop/src/Templates/BadOpcode
ask Ori to add this opcode once the realisation is certain
| balance: 0 | ||
| # 1025 push0's | ||
| code: ":raw 0x\ | ||
| 5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f5f\ |
There was a problem hiding this comment.
what are you trying to break by calling it 1025 times ?
is there a test cases shit?
I would check push0 in staticcall to see that it act as normal pushes.
also make sure it does not mess up stack if there is some stack pushed already and then push0 happens.
There was a problem hiding this comment.
what are you trying to break by calling it 1025 times ?
This case comes from https://eips.ethereum.org/EIPS/eip-3855#test-cases
I think it checks that exactly 1 item is pushed on stack, so 1025 of them overflow stack.
There was a problem hiding this comment.
also make sure it does not mess up stack if there is some stack pushed already and then push0 happens.
how would you check this?
There was a problem hiding this comment.
also make sure it does not mess up stack if there is some stack pushed already and then push0 happens.
how would you check this?
init:
store: { 01 : 10, 00 : 10 }
code: 60015f55
after the exuction the corresponding cell must change to 0
| 0000000000000000000000000000000000000100: | ||
| balance: 0 | ||
| # push1(01) push0 sstore | ||
| code: ':raw 0x60015f55' |
There was a problem hiding this comment.
would that also obsolite "0x6000"?
There was a problem hiding this comment.
No, PUSH1(0) is still valid
|
The work continues in another PR #1033 |
https://eips.ethereum.org/EIPS/eip-3855
Geth implementation in ethereum/go-ethereum#24039