Hi everyone,
After working on several Rust codebases (both personal projects and production work), I’ve put together a practical checklist that helps me keep repositories clean, deterministic, and easier to maintain over time.
I’d like to share my current development philosophy, especially around production-grade practices and enforcing quality early.
1. Readability & Self-Documenting Code
- I prefer idiomatic functional style with iterators (
.map(),.filter(),.collect()etc.) whenever it keeps the logic clear and declarative. - I try to make code self-explanatory through good naming and small, focused functions. I use inline comments sparingly — mainly for
unsafeblocks, complex algorithms, or “why” decisions. - Public APIs get proper
///documentation for clean rustdoc.
2. Robust Error Handling
- I treat
.unwrap()/.expect()as code smells in core logic. Panics in production are usually a sign that error cases weren’t properly modeled. - I rely heavily on the
?operator with explicitResulttypes, making error paths first-class citizens in both code and tests.
3. Testing Strategy
- Unit tests live close to the code (
#[cfg(test)] mod tests) for fast feedback. - More complex integration and polyglot boundary tests go into the
tests/directory to simulate real usage.
4. Strict Local Quality Gates (My Favorite Part)
Before every commit, I run a local hook suite:
cargo fmt --checkcargo clippy --all-targets --all-features -- -D warnings(treat warnings as errors)- Full test suite (
cargo test) - Dependency hygiene checks
This gives me high confidence that broken or low-quality code doesn’t even reach the repository.
I've recently been trying to summarize my methods into a skills.md file so that AI can understand it and use it in real-world projects. If possible, please give me some suggestions.
How can style rules be input into LLMs without causing hallucinations?