Monday 6 February 2017
What Rust Can Do That Other Languages Can't, In Six Short Lines
struct X {
y: Y
}
impl X {
fn y(&self) -> &Y { &self.y }
}
This defines an aggregate type containing a field y of type Y directly (not as a separate heap object). Then we define a getter method that returns the field by reference. Crucially, the Rust compiler will verify that all callers of the getter prevent the returned reference from outliving the X object. It compiles down to what you'd expect a C compiler to produce (pointer addition) with no space or time overhead.
As far as I know, no other remotely mainstream language can express this, for various reasons. C/C++ can't do it because the compiler does not perform the lifetime checks. (The C++ Core Guidelines lifetime checking proposal proposes such checks, but it's incomplete and hasn't received any updates for over a year.) Most other languages simply prevent you from giving away an interior reference, or require y to refer to a distinct heap object from the X.
This is my go-to example every time someone suggests that modern C++ is just as suitable for safe systems programming as Rust.
Update As I half-expected, people did turn up a couple of non-toy languages that can handle this example. D has some special-casing for this case, though its existing safety checks are limited and easily circumvented accidentally or deliberately. (Those checks are in the process of being expanded, though it's hard to say where D will end up.) Go can also do this, because its GC supports interior pointers. That's nice, though you're still buying into the fundamental tradeoffs of GC: some combination of increased memory usage, reduced throughput, and pauses. Plus, relying on GC means it's nearly impossible to handle the "replace a C library" use case.
Comments