Closed holiman closed 3 months ago
It's fine btw if you decide not to fix this, I don't usually fuzz older forks, so I can just keep to fuzzing post-merge forks where this is not a problem
Getting old forks correct is pretty important for us!
@holiman I see from the trace comparisons that eels
(eelsbatch-0) is loading 0x200000
for the DIFFICULTY
which is also the value provided by currentDifficulty
in the env
"currentDifficulty": "0x200000"
Right! The flaw is with the other clients! Hah!
When
eels
excecutsDIFFICULTY
on aLondon
fork, it takes theenv.currentRandom
value from the statetest, notenv.currentDifficulty
. See difference in stack contents:Example testcase: