Add unsafe_op_in_unsafe_fn to Clippy CI & Crates

[bushrat011899]

Objective

• Fixes Lint against (and fix) missing unsafe blocks in unsafe functions #9848

Solution

• Added unsafe_op_in_unsafe_fn to the lint options for CI clippy task.
• Added #![forbid(unsafe_op_in_unsafe_fn)] to all crates that did not already forbid unsafe
• Placed unsafe { ... } statements around specific uses of unsafe where lint raised warnings.

Notes

• My safety comments are fairly generic due to the shear volume required. I would encourage people to submit change suggestions to better comments where they may have a better understanding of the specific unsafe function contract.
• I've neglected the generated.rs file in bevy_mikktspace, as something like Refactor bevy_mikktspace to entirely safe rust #9050 will resolve that issue on its own.

[joseph-gio]

IMO it would be better to add this lint to each crate individually, rather than putting it in CI. That would provide a better development experience since contributors won't be surprised by CI failing when clippy passed locally.

It would also make it easier to fix all of the occurrences since we could do it for each crate/groups of crates separately.

[bushrat011899]

IMO it would be better to add this lint to each crate individually

Yeah I think you're right, I'll add that but keep it in the CI as a catch for future crates.

[Kanabenki]

Once #10011 is merged, this could be rebased to integrate the unsafe_op_in_unsafe_fn in the Cargo.toml lint table 🙂

[alice-i-cecile]

Yeah, IMO we should move this to the universal lint config.

[mockersf]

Could you keep the lint at the warn level rather than forbid? It gives a much better local dev experience, and will be denied by CI

[bushrat011899]

I'm going to have another crack at this now that #10011 is merged, with the goal being to warn on unsafe_op_in_unsafe_fn, and fix any violations. I'm switching this to draft just to indicate it's not ready yet.

[BD103]

There's been some recent work in this area through #11590. Can any of this PR help with that? (Especially bevy_ecs. It would be a shame for all this work to be repeated.)