OSDev.org

I'm happy to announce the release of Sortix 1.0. This is the first self-hosting and installable release.

Sortix is a small self-hosting operating-system aiming to be a clean and modern POSIX implementation. It is a hobbyist operating system written from scratch with its own base system, including kernel and standard library, as well as ports of third party software.

There are a lot of major developments this release:

* Interactive installer.
* Manual pages for Sortix specifics.
* Relicensed to ISC.
* User-space converted to C.
* Custom partition editor.
* Harddisk drivers.
* Init system overhaul.
* And a lot more.

This goal of this release is for Sortix to become a solid base for future growth.

Installation and More Information

See the official website for general information about the Sortix project.

The 1.0 release page goes into depth about this release.

You can download Sortix from your local mirror.

The installation(7) manual page covers the installation process in depth.

The source code is freely available and included in the release itself.

Thanks

I'd like to thank all my contributors that contributed code, tested the release, and gave advice.

Sortix is one of the most advanced hobbyist operating systems I've used. It's great to hear 1.0 version and that I can run it on real hardware.

Your website reads: "169k lines of source code." How many of them are written by you? Do they include ported programs' source code?

And do you have any plans for networking in the future releases?

@sortie, I really appreciate your emphasis on a stable release. In general, with the new license and interactive installer, this project is starting to be more and more interesting. I installed an RC version a week ago on real hardware and everything seemed to work. I will try to contribute a Finnish keyboard layout.

It is important to release something that is usable instead of constantly pushing forward and thinking "I will make a release after it is ready" because it will never happen. Now we have something that clearly marks a point in the Sortix history. Without doing these releases, it is just a fuzzy "another" project without any realistically achievable goals.

Please continue, you are on the right track!

Awesome sortie, I am truly impressed! Installation worked flawlessly, great work.

Hi,


Nice work! It's very positive that there is such an improvement over 0.9. I like it.

I'll also try to contribute more for the next releases.

Muazzam wrote:Your website reads: "169k lines of source code." How many of them are written by you? Do they include ported programs' source code?

This is a fair question, though consider how much software has been ported to Sortix. GCC alone is more than 169k lines of code. So it's evident that these 169k lines of code were written by sortie, except the touch(1) utility which was actually written by me.


Regards,
glauxosdever

Congrats!

I tried the release and it seems stable as I haven't had a crash yet.

-- sinetek

Thanks guys, it really means a lot to hear your comments.

Great to hear it works and is stable!

Muazzam, the 169k lines (according to wc -l) only counts the base system, which is written by me, with the exception of the netbsd libm which is 28k lines. It doesn't cover the ports collection, which is really massive, thanks to ports such as gcc.

Networking is something I'd like to work perhaps for the next release or the one after that. I also have to work on many more drivers like CD-ROM, USB, and so on. There's no shortage of interesting work, and I'll see what I manage to do. Networking is however a feature I'd like to prioritize, as it enables a lot of stuff and would increase the fame of Sortix. A contributor already did some initial work on this, which I look forward to looking into.

Antti, thanks, I welcome contributions. Keyboard layouts seems to be a thing several people want to contribute, and a fairly simple way to get started, and a way to make Sortix theirs, so I'm absolutely accepting such patches. Stable releases are important, and I'd love to see even more projects do those. They are statements that "here is my project, and it's something to be proud of". I've made a habit of doing them and incrementing by 0.1 each time, hence 1.0. This release got a bit delayed, I was meaning to get it out by the end of 2015, and we'll see when it fits me to do the next release. Probably between 8 to 12 months from now.

I've set myself up nicely with 1.0. It's easy to do a new release whenever I want, as the system is stable, if I avoid regressions. It's easy to upgrade to newer releases, and my nightly build server does build such releases of master. I can just work for whatever amount of time I want, and when I feel sufficiently better than 1.0 or enough time passes, I can just do a new release. This stuff matters and sets yourself up for success in the real world.

It was fun announcing it on twitter also, where I got three dozen retweets. There's interest in this sort of thing if you can show it to the right people. I even got a couple unsolicited contributions already.

Thanks! <3

I've also managed to port some software to it (the silver searcher)

sortie wrote: Muazzam, the 169k lines (according to wc -l) only counts the base system, which is written by me, with the exception of the netbsd libm which is 28k lines. It doesn't cover the ports collection, which is really massive, thanks to ports such as gcc.

Working on my OS almost daily for nearly a year, I could only produce ~7,000 lines of code. You must be orders-of-magnitude better programmer than I am!

Muazzam wrote:Working on my OS almost daily for nearly a year, I could only produce ~7,000 lines of code. You must be orders-of-magnitude better programmer than I am!

Lines of code are not related to the quality of the software. Maybe the extent, but not the quality.

max wrote:

Muazzam wrote:Working on my OS almost daily for nearly a year, I could only produce ~7,000 lines of code. You must be orders-of-magnitude better programmer than I am!

Lines of code are not related to the quality of the software. Maybe the extent, but not the quality.

Sure. But it can be a good estimate. And I can trust Sortie that all (or 99%?) of these 150k+ lines would be useful.

There's some bloat in my tree still, but it's not that bad. It's a process. And Sortix definitely isn't full of quality code, some of it is fairly decent, but worse is sometimes better. As I age, I start hating older Sortix code more and more, because I get higher standards.

Muazzam, I've also been programming for well over a decade and worked on my operating system for five years, and attend university. Still, I could output a lot more code if I wanted to, the bottleneck is not writing code, but all the rest, including doing it well and deliberately not writing code (no code is correct code). The first 7000 lines of an operating system is probably the hardest, as you have to get well started, after that it's more of the same old. Just keep working well at it and you can become orders of magnitudes better. (That said, I predict 2026-sortie to be scary.)

On the other hand, some of the best programmers produce negative several thousand lines on a good day. It's lines spent, not written.

sinetek, that's just amazing. Wonderful seeing what people do with it.

Congrats! These are amazing news! Just installed it in VirtualBox. By the way, why does killing the init process results in some strange behavior? Shouldn't it just shutdown or refuse to be killed?

Roman wrote:Congrats! These are amazing news! Just installed it in VirtualBox. By the way, why does killing the init process results in some strange behavior? Shouldn't it just shutdown or refuse to be killed?

POSIX doesn't specify behaviour of init, so it's entirely implementation defined. It *should* do something special, but it's not required or even suggested anywhere in POSIX as far as I know.

It would be a good idea, though. Something for 1.1, I guess.

I've tried it and it works flawlessly! BTW, a Bochs port is really cool!