• Hexarei@programming.dev
    link
    fedilink
    arrow-up
    0
    ·
    10 months ago

    I beg you forgive my pedantic interjection, but … I posit that the original commenter is incorrect. it is absolutely native execution.

    The CPU is fetching and executing the instructions directly from memory, without any (additional) interpretation of code or emulation of missing instructions - Which is, by definition, native execution.

    What the compatibility layer “does” is provide a mapping of Windows system calls into the appropriate Linux system calls. Or, in other words, makes it so that calls to functions like CreateWindowEx() in the Win32 API have a (still native) execution path.

    The native execution requires you to install WINE, yes, but if we’re disqualifying it because “it requires you to install a package”, then we also consequently:

    • Add things like “print stuff”, “display graphical applications”, and “play audio” to the list of “things Linux can’t do”
    • Disqualifies Windows from “natively executing” any .NET applications (a Microsoft-built first-party framework), since .NET applications require you to install .NET.
      • Hexarei@programming.dev
        link
        fedilink
        arrow-up
        0
        ·
        10 months ago

        You’re correct in that it is a compatibility layer - And I’m not disagreeing with that. Also to be clear: Not just arguing to argue or trying to start a fight, mind you. I just find this to be an interesting topic of discussion. If you don’t find it to be a fun thought experiment, feel free to shoo me away and I’ll apologize and leave it alone.

        That said, we appear to only be arguing semantics - Specifically around “native” having multiple contextual definitions:

        • I am using ‘native’ to mean “the instructions are executed directly by the CPU, rather than through interpretation or emulation” … which WINE definitely enables for Windows executables running on Linux. It’s the reason why Proton/DXVK enables gaming with largely equal (and sometimes faster) performance: There is no interception of execution, there is simply provision of API endpoints. Much like creating a symlink in a directory where something expects it to be: tricking it into thinking the thing(s) it needs are where it expects them to be.

        • However, you are using ‘native’ to mean “within the environment intended by the developer”, and if that’s the agreed definition then you’re correct.

        That’s where this becomes an interesting thought experiment to me. It hits me as a very subjective definition for “native”, since “within the intended environment” could mean a lot of things.

        • Is that just ‘within a system that provides an implementation of the Win32 API’? If so, WINE passes that test.
        • If I provide an older/fixed/patched version of a DLL (by just placing it in the same directory) to fix an issue caused by a breaking change to a program that is running on Windows, is that no longer native?
        • Or is it just ultimately that the machine must run the NT kernel, since that’s where the developer intended for it to run?

        Does that make sense? I hear a statement like that and I find myself wondering Which layer along the chain makes it “native”? - I find myself curious at what point the definition changes, in a “Ship of Theseus” kind of way.

        It seems to me that if we agree that the above means “running in WINE is not native”, then we must also agree that “anything written running for .NET (or any other framework, really) is not native”, since .NET apps are written for the .NET framework (Which is not only officially available for Windows, mind you) and often don’t include anything truly Windows-specific. Ultimately, both are providing natively-executed instructions that just translate API calls to the appropriate system calls under the hood.

        I hope that does a better job of characterizing what I meant.

        • JungleJim@sh.itjust.works
          link
          fedilink
          arrow-up
          0
          ·
          10 months ago

          You clearly know more about this than I do, and you’ve thought a lot about it. Your points deserve a better response than I can give at this time, but I wanted to acknowledge that at least. I also wanted to say you aren’t pedantic and I’m sorry I said that. You spent time and thought on making a good conversation and I wish I had been more engaging with that instead of trying to be correct. Thank you for still conversing instead of arguing even after I was less than perfect of a conversation partner. I hope in the future I see more of your comments. Have a really nice day.

          • Hexarei@programming.dev
            link
            fedilink
            arrow-up
            1
            ·
            10 months ago

            I appreciate your acknowledgement - and I commend the humility it takes to write a comment like this! No hard feelings at all, and I hope things are pleasant for you as well.

            It’s folks like you and interactions like this that make Lemmy a platform worth engaging on.