Announcement

Collapse
No announcement yet.

Rust 1.0 Now In Beta

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Rust 1.0 Now In Beta

    Phoronix: Rust 1.0 Now In Beta

    The beta of Rust 1.0 was released ahead of Easter weekend and all libraries/language features planned for the v1.0 milestone have been marked as stable...

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    "Stability" in this sense is the same as forwards compatibility. That is to say, a Rust 1.0 source file should compile under a compliment Rust 1.3 compiler, for example. I hope stability accelerates the adaptation of Rust as a gcc front end.

    Comment


    • #3
      Originally posted by risingganymede View Post
      "Stability" in this sense is the same as forwards compatibility. That is to say, a Rust 1.0 source file should compile under a compliment Rust 1.3 compiler, for example. I hope stability accelerates the adaptation of Rust as a gcc front end.
      Why would you need a gcc frontend?

      Comment


      • #4
        Originally posted by Szzz View Post
        Why would you need a gcc frontend?
        Something I've learnt the hard way is that you always want multiple competing implementations of a language. The main benefit of doing so is that you end up with a better language spec, since the different implementations tend to identify ambiguity in the spec, instead of something that in practice depends heavily on the quirks of a given implementation. There's also the practical aspects of having an alternative in case a project dies, and being able to compare the performance of the different implementations. (Not to mention that GCC still has better support for obscure architectures than LLVM.)

        Comment


        • #5
          Any chance we could see the Linux kernel be rewritten with Rust...say version 2.0+, within the next decade?

          Comment


          • #6
            Originally posted by Krysto View Post
            Any chance we could see the Linux kernel be rewritten with Rust...say version 2.0+, within the next decade?
            None whatsoever. The kernel is still written in C89, and I'd say the chance of it moving to C99 within the next decade is maybe 60%.

            Even if that weren't the case, I've worked with Rust enough to realise that porting C code to it is significantly harder than just writing it from scratch, because of the constraints imposed by its memory model. Besides, Rust-based kernels have already existed for a while, so you could always use one of those (or even better, seL4, which has been formally proven to be functionally correct) instead.

            Comment


            • #7
              Originally posted by Krysto View Post
              Any chance we could see the Linux kernel be rewritten with Rust...say version 2.0+, within the next decade?
              Almost certainly not the kernel, but perhaps modules, libraries, and gnu-utils (which has been ongoing for many months).
              As rdnetto says, it makes more sense to use something like seL4 which is already great, but has very limited functionality next to a gp kernel. Using rust, however, you could bootstrap the remaining functionality as servers/daemons/whatever the term they use and have a pretty damn safe system.
              The big issue is, as always, drivers. seL4 had a plan for that but I've not read any updates for awhile (involved auto-negotiated drivers from schematics http://ssrg.nicta.com.au/projects/TS/drivers/synthesis/ and constraints).
              All that being said, we need to keep the IoT in mind, and, I'm afraid, mbed is really making some noise (http://www.eetimes.com/document.asp?doc_id=1326223). I'm not sure how difficult it would be to migrate their power management apis but they are worth looking at, imho.

              Comment


              • #8
                Originally posted by risingganymede View Post
                "Stability" in this sense is the same as forwards compatibility. That is to say, a Rust 1.0 source file should compile under a compliment Rust 1.3 compiler, for example. I hope stability accelerates the adaptation of Rust as a gcc front end.
                not a chance if it is not backwards compatible to previous several decades of existing code

                Comment

                Working...
                X