The reason I personally am silent about it is because there is an ongoing overhaul of standard io libraries. Honestly, my need to be adventurous dried up when I was left with few broken libs. Writing about Rust still has a risk of becoming obsolete and misleading quickly. Waiting for the real 1.0.
Strikes me as simply a very appropriate use of macros. Get tired of writing the same syntactic fragment again and again? Write a macro. Want to see what some macro is "hiding"? Look it up or expand it.
At the moment it's just return and try!, but people look to the standard library for what is acceptable. When the standard library contains a macro that can return, people will write their own macros that return. It could potentially be half a dozen different macros you need to keep in your head.
Personally, I go back and forth on it. Hopefully it will turn out fine.
What other solutions are there? The only other approach to error handling I've seen is exceptions (e.g., C++, Java, C#, JS…), and if you don't like `try!` because it is a "hidden return", you certainly won't like exceptions. At least with Rust's macros, I know that in the absence of one, there is no return; in the presence, there might be. Exceptions in most languages make no guarantee.
Is this because the behavior is sort of "exceptional" but not so exceptional that the (supposedly inefficient) exception-mechanism is warranted?
In that case, I think the compiler should handle this case still. Using profiling, it could determine which exceptions are really exceptions and which ones are not.
The broad use of case statements leads to one more odd problem - knowing when, and when not, to use a `;`. Explicit returns are frowned upon, they prefer the "results from the last expression" form of returns. The `;` results in an expression returning a different value and a different type. The type system will usually catch these errors and print a helpful "perhaps you should remove the ';' from this line" message, but it's an extra bit of cognitive overhead induced by case statements.
Ultimately, I think it's less about missing syntactic sugar, and more about the type system acting like an electric fence instead of a hedge in its efforts to guide the user to their destination.
That's not true in any of the code I write. I prefer explicit returns in all of my Rust code. The only time I use the "result from last expression" return is when it's the last statement in the function.
The recommended way to deal with errors in Rust is "try!". Using "try!" essentially gives you the ergonomics of exceptions. You should prefer that to match or .and_then(), which are verbose.
fn foo() -> bool {
true
}
is preferred over fn foo() -> bool {
return true
}
But that's more of a style issue.As for `;`, it's just like Standard ML. `;` is for sequencing expressions. I love it.
Rust is going to be very important. The ownership system is a major step forward in language design. It's a huge improvement over C/C++.
It's not easier to write than C++. Rust may feel clunky for people coming from Javascript, Python, Ruby, and PHP. Having to think about lifetime issues for mere strings is a new cognitive load. The big win with Rust is that most of the errors are caught at compile time. This is Rust's big advantage, but alien to scripting programmers. The Rust compiler report errors in three phases. First you get all the syntax errors, and until the syntax is perfect, that's all you get. Then you get all the type errors, and until the type issues are perfect, that's all you get. Then you get the ownership errors. Ownership is a global analysis; ownership problems involve at least two points in the program. The compiler produces good, but very wordy error messages. (Hint to Rust developers: put in a line length limit and word wrap for long compile time error messages.) If your ownership design is faulty, the result is likely to be "fighting with the borrow checker", because the problem isn't local, and just fixing the compiler-reported error will make the problem pop up elsewhere. The cleverness of the ownership system is impressive, but some programmers are going to feel like they're being hammered by it. Successful C++ programmers won't have a major problem with this. It may be tough on the Javascript crowd.
Rust requires some advance planning, which may be incompatible with "agile" development. It's also difficult to port code from other languages to Rust without rethinking the ownership and bounds logic. There is a port of Doom to Rust. It has a lot of unsafe code, because Doom's internal memory structures are not directly compatible with Rust's. Such problems will recur as big packages with delicate internals are ported over to Rust.
This is partly a documentation problem. The current tutorial (http://doc.rust-lang.org/book/hello-cargo.html) is relentlessly upbeat and glosses over too many of the hard problems. Once some third-party books have been written, that situation should improve.
Every time I've fought with the borrow checker, it's because I've had a serious design or conceptual flaw. (Well, apart from syntax/compiler questions). Since you can derive most of the rules just by thinking about it, I find it grows on you quickly. I'll be very saddened if the borrow checker actually ends up being hurtful for adoption overall. Though I agree if you cannot handle pointers or think about memory, Rust will be difficult. So yeah, scripting only devs will have trouble. But! It's better than them writing the code in C.
Rust overall seems like that. There's less random stuff and things work mostly by thinking about safe, zero overhead abstractions, and what falls out from those mandates. Mostly.
Now, maybe if I was a modern C++ programmer, I'd find the effort about the same. OTOH, C++ systems don't end up as safe as Rust ones, so I'm not sure there's a perfect comparison to be had. Maybe that'll change as C++ has started catching up feature wise, but someone I doubt it.
Writability is a worthy thing to improve: GHC's -fdefer-type-errors feature is a good example. I hope Rust improves its writability in the future.
Around 2002, when there was concern about computer-related terrorism, I suggested that the C++ standards committee's unwillingness to deal with memory safety constituted material support of terrorism. They were angry, and terrified. That post was actually deleted from their USENET group.
1. A tiny function which loads textures. I was fooling around with optimising load speeds. The actual speedup I got was insignificant and I really should revert it to safe code.
2. A silly getter on the transformation matrix---another misguided attempt at a speedup; no reason to use UnsafeCell instead of RefCell. Should revert.
3. Casting buffers from files as structs. This is safe for any buffer the same size as a Copy struct, std offers no such function, so I do. Unsafety is isolated in two small functions (one for T and one for Vec<T>), all its callers are safe.
4. Interacting with OpenGL. Fair enough, this happens all over the place, but it's not actually unsafe. The OpenGL bindings library didn't take much care to only mark unsafe operations as such, and all GL calls are marekd unsafe---which is why I wrap them up in a macro which peforms the operation in an unsafe block and panics on any error. I should port my code to glium and then this would go away as well.
*better formatting
> In particular, allocating a new object and returning a reference to it it from a function is common in C++ but difficult in Rust, because the function doing the allocation doesn't know the expected lifetime of what it returns.
This is what boxes are for. A Box is a unique pointer to a value on the heap and can be used without knowing compile-time lifetimes. References and lifetimes allow you to safely return pointers to stack allocated objects. In C++, you'd have to do this:
MyType value;
my_function(&value);
When returning references, rust uses the lifetimes instead of explicit declarations to figure out where (on the stack) `value` needs to be allocated.> Declarations are comparable in wordiness to C++.
Only at interfaces where the declaration also serves as documentation. Elsewhere, types can generally be inferred.
> Rust has very powerful compile-time programming; there's a regular expression compiler that runs at compile time. I'm concerned that Rust is starting out at the cruft level it took C++ 20 years to achieve. I shudder to think of what things will be like once the Boost crowd discovers Rust.
Unlike C++, 1. Macros from one crate aren't imported into another unless the user explicitly requests that they be. 2. Macro invocations are clearly macro invocations. You never have to wonder if something is a function or a macro.
> The lack of exception handing in Rust forces program design into a form where many functions return "Result" or "Some", which are generic enumeration/variant record types. These must be instantiated with the actual return type. As a result, a rather high percentage of functions in Rust seem to involve generics.
How is this a problem?
> There are some rather tortured functional programming forms used to handle errors, such as ".and_then(lambda)". Doing N things in succession, each of which can generate an error, is either verbose (match statement) or obscure ("and_then()"). You get to pick. Or you can just use ".unwrap()", which extracts the value from a Some form and makes a failure fatal.
I agree that this is less than ideal. However, IMHO, this is better than Java and C++.
Java:
Libraries tend to bubble everything. This leads to long throws clauses in function signatures with unexpected exceptions. A user of these libraries often catches and ignores these exceptions when writing the first draft of his or her programs because they don't make sense (why handle IO Errors when using a collection?). And then, because his or her program works, he or she forget about the ignored exception cases turning them into silent errors.
On the other hand, in rust, you can only return one error. When writing a function that has multiple failure modes, this forces the programmer to think about the set of failures that can happen and come up with new error type. This doesn't force the programmer to come up with a meaningful error type but it gives them the opportunity.
Additionally, like in Java, Rust programmers can ignore errors (`unwrap()`). However, unlike in Java, these ignored errors are not silent, they are fatal.
C++:
Exceptions are unchecked and everyone I've talked to avoids them like the plague. In the end, C++ exceptions end up acting like rust's `panic!()` because programmers don't check them but are used like Java's exceptions because programmers could check them.
> There's a macro called "try!(e)", which, if e returns a None value, returns from the enclosing function via a return you can't see in the source code. Such hidden returns are troubling.
I agree that hidden returns can be troubling. However, in rust, only macros can lead to hidden returns, macros use a special syntax (`macro_name!(args...)`, and macros have to be explicitly imported.
> All lambdas are closures (this may change), and closures are not plain functions. They can only be passed to functions which accept suitable generic parameters. This is because the closure lifetime has to be decided at compile time.
The first sentence is correct but the last two are just wrong:
fn takes_a_function(f: Box<Fn()>) {
(f)();
}
fn main() {
takes_a_function(Box::new(move || { println!("hello world") }));
}
The `Box` allocates the closure on the heap and the `move` causes the closure to capture by value. This means that this closure (`f`) can be moved freely without lifetime restrictions because it doesn't reference the stack. However, most functions that accept closures use generics and do any necessary boxing internally to make the user's life easier.> Rust has to do a lot of things in somewhat painful ways because the underlying memory model is quite simple. This is one of those things which will confuse programmers coming from garbage-collected languages. Rust will catch their errors, and the compiler diagnostics are quite good. Rust may exceed the pain threshold of some programmers, though.
Rust is a systems language. It exposes a lower-level (not simple) memory model because systems programmers need it. If you want garbage collection, you are free to roll your own (yes, you can actually do this in rust).
> Despite the claims in the Rust pre-alpha announcement of language definition stability, the language changes enough every week or so to break existing programs.
Re-read those claims. Alpha means fewer breaking changes and no "major" breaking changes not stability.
OMG, thank you for including this. I spent several months reading every bit of documentation that was available for Rust, and programming in it daily. Made some good progress. But I never, never came across this explanation. Very enlightening.
Rust desperately needs documentation covering these kinds of details. How on earth is someone supposed to make serious use of the language without knowing this?
I believe the Klabnik documentation hinted at this (something like, "The Rust compiler is smarter than that" and therefore you don't need to overuse pointers), but by no means did it actually spell it out. And you only needed a few sentences to cover it.
I know the Rust community is aware that more documentation is needed and has a todo list a mile long. But I don't know if technical details such as this are high enough on the priority list.
There are actual features which still have no documentation. It's hard being a single person trying to keep up with changes from tons of other people, many full time and some community. I may be the person who is most looking forward to Rust being stable...
Given the reference to Boost, the author is almost certainly talking about template metaprogramming, not C macros. TMP is obviously a lot more limited in scope than Rust macros, but it could hardly be called dangerous; I doubt anyone's ever invoked it by accident.
This is explicitly called out as non-idiomatic behavior in the documentation, however. The preferred action is to allocate on the caller's heap and pass a mutable reference down to the callee.
In fact, in general it's recommended not to use Box, because it complicates human reasoning about the code. And while it gets around a lot of the compiler's restrictions, its akin to writing <language of choice> in Rust, which is frowned upon in any language. Recommending its use so broadly is doing a disservice to people who want to learn Rust.
> Only at interfaces where the declaration also serves as documentation. Elsewhere, types can generally be inferred.
Except where they can't, and those locations aren't terribly consistent. The Rust designers have publicly announced their preference for explicitness over inference, and the language reflects that.
> On the other hand, in rust, you can only return one error.
This is not unique to rust, or any language really. You can only throw one exception at a time. You can only set one errno at a time. You can only return one `error` at a time.
> macros have to be explicitly imported
Except for the built in ones, which are the only ones referenced by the OP. Also, by placing the macro delimiter `!` between the name and the parenthesis, it makes the macro harder to scan for visually. I imagine that any editor will want to set up special rules to highlight these distinctly, and having special highlighting for the ones known to change the program flow would be beneficial.
> It exposes a lower-level (not simple) memory model because systems programmers need it.
Low level memory is simple: write to, read from, write to referenced, read from referenced. The OS adds one more major operation: get heap memory. Everything else is added by languages or libraries.
That said, Rust's restrictions on memory lifetimes results in more simplistic memory related code. When you have to jump through extra hoops to create a pointer which may be used beyond a single scope, and the compiler creates so much friction when you want to do anything with them in that greater scope, people will defer back to simplistic memory code.
I'm not certain if this is good or bad; it just is at this point.
> Alpha means fewer breaking changes and no "major" breaking changes not stability.
Any breaking changes affect stability, affects documentation (Rust's library documentation is behind the actual code as of a week ago), and affect 3rd party libraries. The results of this is that if you're not Mozilla, there are significant barriers to writing Rust code right now, and I would not personally recommend learning or writing Rust right now to anybody.
Why aren't they consistent? The Rust type inference is generally very good, and the places where you have to annotate are places where any typechecker would force you to annotate, because the types are simply underconstrained (e.g. the return type of Vec::collect or mem::transmute).
> The Rust designers have publicly announced their preference for explicitness over inference, and the language reflects that.
As the original author of the typechecker, I can state that the idea that we intentionally made the type inference less powerful than it could have been is totally false. It's always been as powerful as we could make it, except for interface boundaries (where type annotation is needed because of separate compilation anyway).
Care to link/elaborate this? I thought the recommendation was to return by value in this case -- making it easy for the caller to decide where to store the value. The move semantics would then optimize the copy away, so you end up using either the caller's stack or the heap, depending on how the call was made.
Really? I've always understood it was because when possible that decision should be left to the caller and boxing by default just made the interface less flexible/convenient for callers. How does Box complicate reasoning about the code?
I'd like to see a code snippet explaining this problem.
fn make_a_foo() -> Box<Foo> {
Box::new(Foo { a: 5 })
}
If the function allocated memory and only returned a borrowed reference, who would be responsible for freeing it? Yes, Rust will make you stop and think there, as it enforces memory safety.In cases where it does make sense to return a reference to a new object, like allocating from an arena, the lifetime ('a) of the returned reference will be the same as the lifetime of the arena.
fn new_from_arena<'a>(arena: &'a TypedArena<Foo>) -> &'a mut Foo {
arena.alloc(Foo { a: 5 })
}
But Rust can infer the lifetime, so that can be shortened to: fn new_from_arena(arena: &TypedArena<Foo>) -> &mut Foo {
arena.alloc(Foo { a: 5 })
}What? C++11/14 solves these issues.
You're right that C++ provides a solution to the first two, but C++ locking via std::mutex isn't done in the same way as Rust: in Rust the mutex owns the data and prevents you from getting access to it unless you lock. std::mutex, however, is a separate value from the data it protects and it's up to you to coordinate access to that data.
I would also argue that Rust is a better solution to the first two issues. Modern C++ does not solve the problem of use-after-free (dangling references and invalid iterators are very possible, and common in large codebases). This is something that I don't believe C++ can solve without becoming a radically different language. Furthermore, Rust forces you to use the right patterns unless you type "unsafe": this is, again, important for security, reliability, and developer productivity, reducing the amount of time you spend in the debugger.
Rust compels correctness, so for any project where you can't trust your coworkers aptitude towards correctness (and that is to say you even trust your own) Rust is an insane productivity booster. We could have avoided millions of hours of work and thousands of zero day and system destroying bugs if we had OS cores written in a language like Rust.
Rust’s synchronization primitives are immature — they’ve been rewritten once or twice in the past year or so — but cool from a usability perspective.
edit: oh, hello pcwalton. I suspect you knew this already. :P
Rust enforces memory safety at the language level. C++ itself does not "know" about memory safety. This difference, to me, is huge. You can still opt out of memory safety in Rust through unsafe regions, but the fact that Rust provides memory safety guarantees to non-unsafe regions is, to me, a change in kind, not degree. When the only thing enforcing memory safety is disciplined use, it's still too easy to make a mistake.
Sure, it solves some of those problems if you use exclusively smart pointers and vectors, and never use the built-in language pointers and arrays. What forces you to do so, when '*' and "new" and [] are right there?
Rust pushes you towards the right solution by requiring an "unsafe" block if you use raw pointers or arrays. That doesn't prevent you from using them (such as in the implementations of higher-level constructs or FFI calls), but it does hint that they're the wrong solution for everyday programming.
Many C++ codebases out there are still pre C++98.
I like C++, but I don't see the opportunity where to use C++14 outside hobby projects.
As an anecdote, the (very large) project I work on at $DAYJOB is finally moving from C++03 to C++11 in the next few weeks.
[0] - http://blogs.msdn.com/b/vcblog/archive/2014/11/17/c-11-14-17...
Sure, but I’ve encountered some notable exceptions: LLVM projects (C++11), Playstation 4 games (C++11/14), QT5 (C++11) projects, et al.
I've hit iterator invalidation in a new c++11 codebase. Perhaps c++14 offers something to help it that I'm not aware of?
edit: "unsolvable" in the sense of the language making it impossible,
What makes you say this? And what is your prior language experience?
Right there is the problem. There are many programmers for whom it's their day job. Others depend on their code doing something useful and important. But they're not theoreticians, don't have advanced degrees in CS, and haven't written in a dozen languages. A new language has to be usable by them to get traction.
The Rust community, at this point, is mostly people who know several programming languages and want to try a new one. Look at the comments above from people who compare type systems in different languages and are aware of the strengths and weaknesses of different approaches. Note the references to obscure languages and research papers most programmers have never heard of, let alone used or read. This is not the target market for a new language, if it is to be a success. It has to be used by people who don't debate language theory issues while the Super Bowl is on.
If i had a pound for every time i'd heard this sentiment, i'd be a rich man!