How (not) to use unsafe code

[gbutler69]

@BurntSushi

With respect to surfacing the visibility of unsafe, I don’t know what the right answer there is. It could be anything from automated tools that report metrics to just developing a stronger culture around auditing code manually, or perhaps something in-between.

Could a proposal similar to the following (previously posted to another thread in internals) help in this regard: [Pre-RFC] Another take at clarifying `unsafe` semantics - #31 by gbutler - Rust Internals