Doc: remove a "safety note" made obsolete by dropck for TypedArena

[SimonSapin]

https://botbot.me/mozilla/rust-internals/2015-04-10/?msg=36316959&page=6

[rust-highfive]

Thanks for the pull request, and welcome! The Rust team is excited to review your changes, and you should hear from @brson (or someone else) soon.

If any changes to this PR are deemed necessary, please add them as extra commits. This ensures that the reviewer can see what has changed since they last reviewed the code. The way Github handles out-of-date commits, this should also make it reasonably obvious what issues have or haven't been addressed. Large or tricky changes may require several passes of review and changes.

Please see CONTRIBUTING.md for more information.

[pnkfelix]

@bors r+ rollup

[bors]

📌 Commit c2fa1f7 has been approved by pnkfelix

[SimonSapin]

@pnkfelix, thinking about this a bit more, I was wondering if dropck really applied here or if unsoundess was still possible with TypedArena. After all, the lifetime of &mut T references returned by TypedArena::alloc are set artificially with transmute.

So I tried to see if I could deliberately cause a use-after-free in safe code:

#![feature(rustc_private)]externcrate arena;use std::cell::Cell;structFoo<'a>{data:Box<u32>,other:Cell<Option<&'aFoo<'a>>>}impl<'a>DropforFoo<'a>{fndrop(&mutself){let other = self.other.get().unwrap();println!("Accessing other item’s data = {}",*other.data);// Use after free?println!("Dropping with data = {}",*self.data);}}fnmain(){let arena = arena::TypedArena::new();// Create a reference cycle within the arena:let a = arena.alloc(Foo{data:Box::new(42),other:Cell::new(None),});let b = arena.alloc(Foo{data:Box::new(6),other:Cell::new(None)});
a.other.set(Some(b));
b.other.set(Some(a));}

This fails to compile with error: arena does not live long enough, but compiles fine if I remove the Drop impl. So I assume this is dropck doing its job correctly, although the error message does not mention Drop at all. It looks like my arena-tree crate only works because nodes (which have &T cycles within an arena) don’t have a destructor.

Just because I didn’t manage to find a problem doesn’t prove it doesn’t exist, but I’m slightly more confident about removing this warning now.