Java has different compiler implementations not all of which necessarily go through interpreter stage. .NET CoreCLR does not have one in the first place (there is technically interpreter code in the codebase but it has bitrotted and is unused). JIT or not does not make it inherently more prone to miscompilation bugs. One thing that Go has going for its compiler is it is a massively simpler design. The downside of that, of course, is much worse compiler output that cannot cope with complex code and is inadequate at high-performance scenarios, where the users of Go either have to try to work around this with goasm or might not even have a solution at all, short of rewriting the component in a different language and praying that the FFI call overhead is low enough for the workload in question.
In addition to that, runtime implemented in C/C++ vs Java/C# or other language does not make it inherently safer. It is always a lower-level flavour of the language, where it cannot depend on itself, and with likely numerous relaxations in assurances to stay performant that the authors uphold manually. Even when C or C++ are used - it's usually a limited, focused subset of the language. And it is a bold claim that these runtimes are not extensively tested, fuzzed and otherwise verified against defects before going to production. If you look at bug and CVE history of .NET - pretty much all issues happen in complex "non-core" APIs, often at the unmanaged side provided by operating system rather than .NET itself (the cryptography bits). I can't recall any that are directly or indirectly related to a compiler bugs.
I feel like this reply ignores the wider context, which is plain old web dev, where the use of Rust and the decision fatigue that it carries is likely to be misplaced. Just pick C# or Kotlin.
I seriously doubt it. Naively incrementing Ref counters is extremely expensive. Tracing GCs in modern languages are very advanced. The least you could do is try to elide some of the ref increments - but you can't do that in Rust. Also boxing every object is very unpleasant. If you're gonna do that, use a language with reference semantics, like Java or C#
> significant chunks of the runtime are still in C/C++
Means absolutely nothing about the safety of the languages themselves. Just because C# has a runtime written in C++ does not mean that you can now magically compile and run use after frees or whatever.
That said, this is just my intuition here, I've never seen a good benchmark trying to do the comparison, so I don't want to over sell it either. But a lot of times, it's like, "okay I have four threads so I need the refcount to go to four" and then it's just using normal references to actually use the value inside.
Rc is cheap to the point of free unless it frees - it’s literally just an addition or subtraction with no data dependency in the path. Your CPU can do > 6 billion of these per second on a 6Ghz CPU (if I recall correctly you could do 3 per clock cycle in a micro benchmark due to out-of-order but at a minimum you should be able to do 1 per cycle).
Arc is quite expensive, particularly on x86 because of its memory model. You can maybe do a few million of them per second. But as you said, you shouldn’t be incrementing ownership once you’ve sent it to a new thread and instead using references directly or using something Rc-like to track the lifetime (i.e. when the current thread’s Rc count hits 0, release your Arc reference).
Runtime bugs are potentially exploitable, even from within the language implemented by the runtime. But I don't think anyone's suggested that they'd present as simply as you suggest. Many such bugs have been found in runtimes of many languages.
Compiler bugs can be similarly exploitable, with the caveat that any code eliminated at compile time won't be accessible at runtime, so compiled languages will tend to have a smaller surface area for exploitation in general.
As security experts like to repeat, it’s a numbers game. You only have to make a mistake once to result in an insecure system. And C++ encourages you to make those mistakes.
Same with JIT - because the compilation happens at runtime, a common attack mechanism is to exploit a particular bug in the JIT implementation so that the generated code violates some assumption and results in unsafe execution. This happens all the time with V8 and a large part of that problem comes from having to implement the language twice (once in the interpreter and once in the JIT and any deviation between the two implementations is an opportunity to exploit a vulnerability). This isn’t something I made up. This is from talking to people who are familiar with this and reading analyses by people who know what they’re talking about.
Compilers btw also have miscompilation bugs all over the place. The reason it’s less of a problem is because the code being executed isn’t “malicious” - it’s your code and you’re not going out of your way to structure code to find bugs in the compiler. This also protects most JIT and why most C# programs don’t have this problem - they’re not loading arbitrary code at runtime to execute. It is a particular problem for V8 though where that’s an expected behavior.
And op is correct. In terms of memory safety, Java and C# aren’t as safe as you might hope because significant chunks of the runtime are still in C/C++ (not true for Go) and because of JIT miscompilation bugs (also not relevant for go). Of course JIT miscompilation in theory should be the same as a compiler miscompilation, but in practice it turns out to be worse because you have to maintain consistency between the interpreted and all the tiers of JIT at runtime and in practice humans have a hard time doing that which is why V8 constantly has security vulnerabilities in this area. That’s why GraalVM is interesting because it proposes a JIT design where you only have the compiler implemented once.