META: FRAME Macro Ideas

[shawntabrizi]

Let this be a collection of nice to have macros and attributes we should add to FRAME.

Where possible, link to bigger issues with more details.

• Syntax improvement for ResultQuery usage #259
• Migrate existing benchmarks! macro to attribute macros substrate#10848
• benchmarking v2 syntax is confusingly lax about return types, leading to confusion substrate#13278
• generation of real benchmark functions for benchmarking v2 substrate#13224
• #[must_use] for benchmarking components #285
• Add special tag to cache specific runtime storage items. substrate#10251
• Vision: Improve Usage of Constants in Pallets #303
• Consider creating a convention to expose the most common Config types for discoverability #308
• Ensure every pallet for which Config is implemented is declared through construct_runtime. substrate#9130
• Production / Non-Production Feature Flags for FRAME Pallets and Runtime substrate#8965
• StorageNMap follow up improvments #321
• Better child storage support #368
• Make procedural documentation compile substrate#1441
• pallet::call macro does not preserve lint attributes #249
• Optional storage hashers in pallet dev mode substrate#13263
• doc-only module to group all storage items of a pallet substrate#12398
• Fix pallet error spam when there is really just one simple error #213
• Encode required origin for extrinsics in metadata #349
• FRAME: Avoid boilerplate weight macros #205

[xlc]

• Encode required origin for extrinsics in metadata #349

[kianenigma]

Another improvement would be to put a stop to this situation: Currently, if your pallet has a syntax error, you get an error for it, and then a gzillion other errors because the corrupt/incomplete pallet code is still being treated as Rust code. Ideally, we'd tell the compiler to stop immediately.

@sam0x17 I am optimistic you'd have a nice hack for this one :)

[sam0x17]

Another improvement would be to put a stop to this situation: Currently, if your pallet has a syntax error, you get an error for it, and then a gzillion other errors because the corrupt/incomplete pallet code is still being treated as Rust code. Ideally, we'd tell the compiler to stop immediately.

@sam0x17 I am optimistic you'd have a nice hack for this one :)

Yes, I actually have some ideas for this. I'll create an issue and add it to the backlog 😄