my source inventory script has found this, i couldn't believe my eyes

Armé el PRIMER tutorial en español sobre @official_fe 🛠️ Si vienes de Solidity, te interesa: vemos setup, tu primer "Hola Mundo" y sus conceptos clave. Míralo acá 👇 youtu.be/f2CL5jAgTKM #FeLang #EVM #Web3

Enter Fe: Both problems collapse into a single compile-time solution. Since bit-width is a const generic parameter, there is zero code duplication. One `StoragePackedArray<BITS>` covers every width we want to support. Because BITS is known at compile time, the slot derivation, lane width, and bitmasks fold into immediate constants. The high-level Fe code compiles down to the same gas as the assembly Solady writes by hand. No assembly needed.

Only eight widths compile: 1, 2, 4, 8, 16, 32, 64, 128. The rule: BITS has to divide 256 evenly. Since 256 is a power of two, those eight are exactly the widths at or below 128 that pack a slot with zero leftover bits. Enforced by a marker trait implemented exactly eight times. Every method carries `where PackedBits<BITS>: ValidPackedBits`, so `StoragePackedArray<7>` is a compile-time type error, not a runtime surprise.

Storage layout is part of the type. `StoragePackedArray<const BITS, const SALT = _>` has a second const generic: `SALT`, the layout seed. The `= _` default tells the compiler to infer it per field. Each field's lanes live in a distinct keccak-derived slot region. For upgradeable contracts that need a stable layout across versions, pin the salt explicitly: `StoragePackedArray<8, core::keccak("v1.tiers")>`.

Solady's LibMap/LibBitMap is beautifully hand-tuned assembly. It also shouldn't have to exist. Every line works around one missing Solidity feature: const generics. In Fe the same thing is ~60 lines, no assembly, same gas. 🧵

不如Fe，我更喜欢rust风格，静态强类型系统。@official_fe

Solady's LibMap/LibBitMap is beautifully hand-tuned assembly. It also shouldn't have to exist. Every line works around one missing Solidity feature: const generics. In Fe the same thing is ~60 lines, no assembly, same gas. 🧵

Bountiful is back. Following the public release of Fe 26.1.0, we've put the new compiler to work on a real-world project that's now live on Ethereum mainnet: a permissionless bug bounty platform with real ETH on the line. 🔗 bountiful.fe-lang.org

Every smart contract dev knows them: the 32-byte magic numbers at the top of files. Storage slots. Type hashes. Selectors. Fancy math. Returned as constants from build scripts, or copy-pasted from somewhere. Taking @Uniswap 's tick math as an example. Fe has a different answer 🧵

hardcoded constants usually hide years of accumulated pain

looks very similar to how comptime in zig works. it's a good idea that allows you to show how the sausage is made and makes the code more auditable, while the bytecode stays is the same optimized shape as if you had hardcoded the magic values.