Ignacio Hagopian

We're also experiencing this issue with the latest version.

> I suggest breaking all the optimizations that are not cow-related into separate PRs, so that we can sift what works and what needs to be reworked. Good work in...

@gballet, I still want to reconsider running this idea again after some other performance improvements we'll include soon. Moving to draft to just avoid noise in the review PR order;...

@gballet, I wonder if this should be closed, considering you mentioned the current spec forces EOAs to store the empty code hash?

Maybe some useful finding: ``` $ readelf -Ws /home/ignacio/.cache/zig/o/569dca02bcfd71a4fcf698e1223e342b/libcompiler_rt.a | grep divei 836: 0000000000000000 1534 FUNC WEAK DEFAULT 476 __udivei4 ``` So `__udivei4` is present (unsigned integer?) but not `__divei4`.

Clarifying here, too, since it might be confusing for others. The same paragraph is in EIP-4762. The link in the PR description can confuse you since, considering it's the same...

> Left some comments. I am not sure I understand why you don't simply use `Get`? When you call `Get`, it will resolve all internal nodes _directly_ in the path...

> As Ignacio said yesterday, we could just introduce a new extension status type but I don't even believe this would be necessary since it's pretty much like adding a...

Dumping here other points mentioned in the call: - The single slot case needs to also include which index is the "unique non-zero storage slot" value in the 256-group. -...

> Now regarding the potential hack: if someone indeed manages to write values that look like an account to a single storage slot (for type 4), then expand it into...