Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 


Forgot your password?
Close
typodupeerror
×
Operating Systems Apache

Rust-Written 'Redox OS' Now Has a Working Web Server (phoronix.com) 53

Posted by EditorDavid from the simple-servers dept.
An anonymous Slashdot reader shared this report from Phoronix: The Redox OS project that is a from scratch open-source operating system written in the Rust programming language now has a working web server, among other improvements achieved during the month of July...

Notable new software work includes getting the Simple HTTP Server running as the first web (HTTP) server for the platform. Simple HTTP Server itself is written in Rust as well. There is also an ongoing effort to bring the Apache HTTP server to Redox OS too.

Another app milestone is the wget program now working on Redox OS. There's also been more work on getting the COSMIC desktop apps working on Redox OS, build system improvements, and other changes.
This discussion has been archived. No new comments can be posted.

Rust-Written 'Redox OS' Now Has a Working Web Server More

Rust-Written 'Redox OS' Now Has a Working Web Server

Comments Filter:

  • This seems trivial. What's the big deal about some hobbyist's OS just because it's written in RUST?

    • Re:....And? (Score:5, Insightful)

      by drinkypoo ( 153816 ) <drink@hyperlogos.org> on Sunday August 04, 2024 @05:00PM (#64680446) Homepage Journal

      I may just be biased because I think they are neat, but I think being written in a "single" language (markup languages, scripts and whatnot aside) is a big bragging right for an operating system. Another one that they are pursuing is self-hosting. Per the announcement they have succeeded in getting some sample programs in Rust, C, and C++ compiled and run inside of their OS, which is a big step. Even if this hobby OS doesn't really go anywhere, it's possibly already taught some valuable lessons, and the dominant user-facing OS on the planet started as a hobby OS.

    • Re: (Score:2)

      by Pembers ( 250842 )

      The announcement of what they did in July says that NLNet has given (or will give) them €26,500, or about US$29,000, which suggests it's rather more than a hobby project.

    • You mean you can't feel the safe?

    • Re:....And? (Score:4, Interesting)

      by caseih ( 160668 ) on Sunday August 04, 2024 @05:05PM (#64680460)

      Except that this is the intended target of Rust in the first place: OS system programming! So if Rust is all that, an OS written in Rust should end up being fast and secure, more than an OS in a traditional language. And should be one of the first things done with Rust. Will be interesting to watch in this space, and also to see what Rust ends up doing in the Linux Kernel. And to see how rust compares in development speed with the existing, less secure languages.

      To me it seems that Rust's biggest negatives are the learning curve and development speed when using Rust. This will improve on both counts as things continue and people become familiar with it and develop best practices.

      It's rather amusing to see all kinds of re-implementations of existing utilities an applications are popping up, their main feature being they are written in rust. Seems like a fad.

      Personally I'm far more interested to see how rust concepts can be adapted by existing languages. For example the Circle compiler for C++. That has immediately utility as one can start using these safety constructs with new code on existing code bases.

      • To me it seems that Rust's biggest negatives are the learning curve and development speed when using Rust.

        I believe Google did an internal study and found that their developers were just as productive in rust as they were golang, so I wouldn't count this as being an issue.

        Personally I have a very hard time going back to golang from it -- rust is definitely a more expressive language.

      • Re: (Score:1)

        by Anonymous Coward

        It's rather amusing to see all kinds of re-implementations of existing utilities an applications are popping up, their main feature being they are written in rust. Seems like a fad.

        Dunno, when I have to learn a new language I like reimplementing something that I've already done in another language. Takes away the "how do I do this" part of the problem so I can focus on the "how do I tell the computer to do this using this language" part of the problem.

        Not only that, but if a new OS were to pop up I am sure

    • Linux started as some hobbyist's OS

    • That's funny. I am sure there were people who were like "This seems trivial. What's the big deal about some hobbyist's OS just because it's open source?" when there was news about Linux in the 90s.

  • We haven't seen any Hurd stories in a while (Score:4, Funny)

    by 93 Escort Wagon ( 326346 ) on Sunday August 04, 2024 @05:44PM (#64680490)

    I guess this new OS can remedy the glaring absence of recent Slashdot stories with titles like "... now works with some printers" or "... now has a working USB driver" - a gap formerly filled by Hurd's maintainers.

    • Slashdot used to regularly post articles about eclectic operating systems. I read about AtheOS, Haiku, DragonFlyBSD, TempleOS, ReactOS and more here. It's news for nerds even if it isn't stuff that matters.

  • when Rust is supposed to be memory-safe? The point of Rust is that it is safe enough to write a monolit while still performing like C and C++.

    • Re: (Score:2, Informative)

      by Anonymous Coward

      Rust doesn't enforce safety because the language allows for unsafe code sections. If the OS is to accept untrusted code from outside developers, microkernel seems like an okay idea.

      • And likely for performance reasons, or for reasons of actually implementing a kernel, they will turn off the safety in some areas of code. Ie, imagine doing mbufs or skbuff (linux) using a strictly safe memory model? Or any kind of buffer pool where ownership is changed often. Then you muck with all sorts of memory mapped device registers. That's really why C plus assembly remains dominant for operating systems, or at least C++ minus features.

  • Can't we move to the future and move beyond UNIX concepts. It's been long enough, there should be a new concept that supports modern hardware better.

    • Re: (Score:2)

      by sgage ( 109086 )

      Good idea - go for it!

    • Re: (Score:3, Insightful)

      by gweihir ( 88907 )

      Not at this time. UNIX concepts are superior to everything else that has been tried. That may change in the future, but not so far. Anything today is either UNIX, UNIX-like or a bad and flawed copy (Windows).

      • Re: (Score:2)

        by cstacy ( 534252 )

        Not at this time. UNIX concepts are superior to everything else that has been tried.

        What means "superior", and just how young and uninformed are you to suppose "everything else that has been tried"?

        • He is not saying every OS that could possibly exist in the future has been tried against Unix. He is saying that Unix has beat every competitor in the past.

        • Re: (Score:2)

          by gweihir ( 88907 )

          Here is a little grammatical lesson for you, as you seem to be especially weak in that regard: "everything else that has been tried" means all the things that have been tried in the past. It does not mean that new things cannot be tried in the future and, in fact, they can be. And they can potentially overcome that limitation I refer to. Oh, and had you read the third sentence I wrote, you would even have seen that I explicitly point that out!

          tl;dr: Learn to read.

      • Re:Can we get an OS where everything isn't a file? (Score:4, Funny)

        by vbdasc ( 146051 ) on Monday August 05, 2024 @01:30AM (#64681044)

        In a tiny OpenVMS voice: Hey, am I a joke to you?

        • Re: (Score:1)

          by jonadab ( 583620 )
          There are things I like about VMS, and one feature in particular that I would dearly like to have (specifically, the way it handles file versioning); nonetheless, realistically, we have to admit that VMS is dead as a platform (though it remains an important piece of computing history).

          • Re: (Score:2)

            by gweihir ( 88907 )

            VMS was really good and ahead back when. Yes, I have worked with it. But it could not keep up with the times and was missing too much stuff in the end.

            • Re: (Score:1)

              by jonadab ( 583620 )
              Yes, that happens when there aren't enough users to motivate the developers to actively maintain the system any more. The chain of buyouts probably hastened its demise (because, with proprietary systems, the company is a single point of failure), but even before DEC got bought, VMS had been in minimal-maintenance mode for at least a decade. The differences between OpenVMS 5 and 6 can be charitably described as "relatively minor".
    • Sure. What is your suggestion and what advantages does it offer over Unix?

    • I don't see what's wrong or outdated about everything being a file, but there's DBOS where everything is a database.

      In a Lisp OS everything is a list and in a Smalltalk OS everything is an object.

      And of course there's Windows where each thing is whatever MS devs felt like that day, often whatever would perform best at the time.

      • The big thing about the Unix idea is to provide a common API for different things, and in this case the API for files is the same as the API for devices, pipes, sockets, etc. Because they all have common abilities: open, close, read, write, set attribute. Then you can find devices via the directory hierarchy, which is really just a small file that describes the device; as opposed to the clunky "A: B: C: D:" style. So really what Unix is doing is making everything uniform, which is a huge advantage.

        I see

    • We can just move beyond files, those crusty old things that kids hate these days... But no, no one's really gotten something newer. They will have applications that try to hide files, such as on a phone, but under the hood they're all just files with a presentation layer.

      And with Unix, the concept isn't that everything is a file-as-in-document, but that everything is a file-handle and all the handles have a common interface whether file or device (or in plan 9 also windows and other things).

      • You could force all applications to store structured data in your doesn't-look-like-a-filesystem database, but at some point you'd have to exchange data with some kind of system that uses files, so you'd have to construct files for interchange.

        But as you probably know, that's how most computer systems worked before Unix, and the idea of using unstructured storage and letting the programs work out handling the data was an advancement in computing. It turned out that it allowed much faster development because

        • I think the Palm Pilot had something different for the application data, but I don't have details handy. Maybe it was a database too.

          • Palm pilot had what are called "library" files but you didn't have to use them. All of the system applications used them for stuff like contacts and reminders so that third party applications could easily manipulate their data files.

    • I agree, files are stupid. We need a folio based OS. Just put everything in folios. Or books. Instead of files put data in books. Organize it with the Dewey decimal system.

  • They still have a better chance of getting self driving working on redox than musk does on a Tesla.
  • Unix kernels use assembly for low level stuff: setting up registers for traps handlers for instance. How RedoxOS manages that?

Slashdot Top Deals

There are never any bugs you haven't found yet.

Close