🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

If this claim is true we should not increase the code size limit.

Lefteris Karapetsas Ledger What's known as pinning in npm (to depend on a specific version) is not a good practice IMO, because it makes it harder for fixes to reach apps: every package in the middle has to opt in to a fix. Ranges and lockfiles are better. Obv loading an unknown version from a CDN is wild.

Approvals leave users overexposed to smart contract risk. We need to move away from them and switch to call- (or time-) bound allowances. If we started from scratch, this is how I'd do it: github.com/frangio/erc690… For existing tokens, a solution like Permit2 can be a workaround.

There is currently no way to make even small security amendments to Final ERCs, and there is no standard place for security advisories about them. We need some place to aggregate these important learnings. I don't want these things to be based on folk knowledge.

Hayden Adams 🦄 Not worth removing events, but if you must, Uniswap v4 could cut down gas spent on events by 70% by simply packing all parameters instead of using the default ABI encoding. Nonstandard encoding will be a pain, but not as much as call traces.

Why do we have "stack too deep" errors? JVM, WASM, CLR are all stack machines too but don't have this issue! Why is EVM different? I looked under the hood at Solidity codegen to find out. Check out my first post in a series about compiling for the EVM: frang.io/blog/codegen-f…

I proposed a small addition to the EIP-712 type system to fulfil a need I've been observing recently: ethereum-magicians.org/t/eip-7713-box…

EIP-7702 or AA on their own are not enough to fix approve UX. We also need to standardize something like EIP-5792. More people should be looking into this!

Having worked on OpenZeppelin Contracts for 6 years I've formed some... thoughts about Solidity. Next week at Devconnect ARG I'll be sharing some of them and what I think might be a good path to explore. I hope you join me! app.devcon.org/schedule/QE7RWH

Pretty surreal to experience the model acting as a faster version of you. Here's an example where o4-mini searched for 1 minute and answered my question by finding and interpreting the relevant source code: chatgpt.com/share/680cfcac…