Phoronix reports that version 0.6 of GNU Hurd has been released. Before getting too excited about GNU Hurd, it's still bound to x86 32-bit and doesn't offer any compelling new features.
GNU Hurd 0.6 has "numerous cleanups and stylistic fixes" to the code-base, the message dispatching code in Hurd servers is now better, there's support for protected payloads of GNU March 1.5+, libz/libz2 are used as the decompressors to replace gz/bz2, the native fakeroot has improved, the performance of the integer hashing library has improved, and the init server has been split into the start-up server and a SysVinit-style program. The procfs and random translators were also merged.
More details on the new GNU Hurd release can be found via the 0.6 release announcement issued by Thomas Schwinge.
(Score: 2, Interesting) by SubiculumHammer on Friday April 17 2015, @05:14PM
When Slashdot ran this news article, the comments were lame. Really lame. A bunch of thoughtless GNU bashing.
(Score: 5, Insightful) by mtrycz on Friday April 17 2015, @05:40PM
I for one am actually pleased with the perseverance of certain initiatives, like Hurd or Minix.
There is no harm really in having more, so people that go GNU bashing are really just bitter dicks. If it doesn't work for you now, doesn't mean that it doesn't work for anyone, or that it can't work in the future. Also, GNU has done just so much for the comunity at large that goes often forgotten that it's straight silly.
Godspeed.
In capitalist America, ads view YOU!
(Score: 2) by frojack on Friday April 17 2015, @07:58PM
There is no harm really in having more,
But isn't this an experiment that has long outlived any possible usefulness? A hobby akin to building sailing ships with match sticks in glass bottles?
Does anybody or anything use these OSes?
No, you are mistaken. I've always had this sig.
(Score: 3, Insightful) by turgid on Friday April 17 2015, @08:15PM
And what is wrong with doing an interesting hobby?
I have some pointless hobby projects too that are 15-20 years behind the times and never get finished (and never will at this rate) but I find them educational and good mental exercise.
You should see my home-made vector font rendering engine that runs on SDL 1.2... :-) Actually, no you shouldn't.
I refuse to engage in a battle of wits with an unarmed opponent [wikipedia.org].
(Score: 3, Insightful) by mmcmonster on Saturday April 18 2015, @12:29AM
The answer is Critical Mass...
Not in the number of individuals using the OS or even developers per se. But once an OS is able to bootstrap itself on modern hardware and run enough hardware drivers so that a desktop can be compiled, then you've made it.
That being said, it's taking a long time for HURD to come along. Long enough that the initial developers could have mated, had children, have them grow up and learn programming, have them help develop the OS, and have them mate and have more kids. :-(
Still, can't wait until they port KDE over. :-)
(Score: 0) by Anonymous Coward on Saturday April 18 2015, @12:47AM
I think it's still worthwhile to examine different approaches to the kernel. because how else can we play around with alternate methods outside of the mainstream kernel?
I still find it interenting that it allows someone to run programs as the user "nobody".
plus it seems like it could have someinteresting possibilities in the VM arena.
(Score: 2) by Common Joe on Saturday April 18 2015, @05:49AM
Everything fails and dies. That will include Windows and Linux and the Mac OS. The only questions are when and how. When the big boys start to die, it is one of these small hobby projects which will take a turn for the more serious and be in the right place at the right time. It won't happen overnight, but it will happen.
I'm not one to experiment with these kinds of OSes, but I am really glad there are people who do.
(Score: 2) by frojack on Saturday April 18 2015, @05:56AM
Fails and Dies?
Nobody has ever woken up one morning to find every instance of a particular operating system inoperable.
Maybe you mean they are replaced by something better? But that's not going to be Hurd.
No, you are mistaken. I've always had this sig.
(Score: 2) by Common Joe on Saturday April 18 2015, @06:37AM
I was speaking generally -- not specifically about Hurd. Also, perhaps "fail" isn't the best word, but "death" is. And as I said earlier, "death" of an O.S. won't happen overnight, but it will happen.
Computers with operating systems / Operating Systems that I have used and are no longer viable in today's world and therefore dead: TRS-80, Commodore 64, DOS, OS/2, Atari, Intellivision.
Yes, I know some of the Atari enthusiast just dug up E.T. and I've seen articles about someone running a C64 somewhere and I've seen DOS used in a few instances of business, but it's safe to say that every single one of these is either is a dead O.S. or a computer that contains a dead O.S.
Apple nearly died a few years ago and is beginning to stumble again without Steve Jobs. Microsoft is not immune from making mistakes (Vista / Windows 8). Linux just is still going through a rough patch with systemd (no matter what your feelings are about the software).
Windows, Linux, and the Mac OS will go away. That's all I was saying. And no, experience tells me that replacements aren't always better.
(Score: 2) by maxwell demon on Saturday April 18 2015, @03:02PM
An operating system turns inoperable when the last piece of hardware it runs on stops working.
Sure, you can revive it by porting it to new hardware, or by writing an emulator for the old hardware. But that's extra work, and unless and until someone puts that work in, the operating system is inoperable, because it cannot run on any existing, working machine.
Indeed, I'm sure there are lots of operating systems that few people ever have heard of, that turned inoperable together with the machines they were written for.
The Tao of math: The numbers you can count are not the real numbers.
(Score: 0) by Anonymous Coward on Friday April 17 2015, @06:53PM
When Slashdot ran this news article, the comments were lame. Really lame. A bunch of thoughtless GNU bashing
You mean hilarious GNU bashing, made even more hilarious because it obviously irritates people like you.
I think GNU HURD is a really good thing. I'm glad people work on it. I'm glad people still believe.
I'm not so glad they're so freaking self unaware that they can't understand how hilarious HURD seems to people outside their community; a little humility goes a long way.
(Score: 1) by SubiculumHammer on Friday April 17 2015, @07:02PM
I barely know what GNU is about, and I hardly got irritated.. I was just asking for our community to show more class than I saw on Slashdot.
(Score: 1, Funny) by Anonymous Coward on Friday April 17 2015, @07:10PM
I was just asking for our community to show more class than I saw on Slashdot.
You could at least try setting the bar a little higher. ;)
(Different AC FWIW.)
(Score: 5, Funny) by Bot on Friday April 17 2015, @07:40PM
They bash GNU because they have no fantasy. A better approach would be like
GNU Hurd:
- X86 only
- decades in development, not even mature beta
- limited hardware support
- still better than having to deal with systemd.
Voila'
Account abandoned.
(Score: 2) by kaszz on Saturday April 18 2015, @12:30AM
Ah, you mean Linux? Just like in the beginning. :P
(Score: -1, Troll) by Anonymous Coward on Saturday April 18 2015, @12:31AM
The GNU Project is still stuck in the 1980s. That's why people ridicule it today. It's almost 30 years behind the curve.
People today don't want restrictive, un-free licenses like the GPL family of licenses. They want licenses like the BSD license and the MIT license that promote true freedom for everyone.
(Score: 2) by maxwell demon on Saturday April 18 2015, @02:54PM
But GNU invented bash! [wikipedia.org]
The Tao of math: The numbers you can count are not the real numbers.
(Score: 5, Interesting) by CRCulver on Friday April 17 2015, @05:35PM
If you are wondering why Hurd continues to be based on GNU Mach, a microkernel that was already regarded as antiquated in the 1990s, this page [gnu.org] has some background. It's disappointing that the attempts to port Hurd to a more advanced microkernel failed. One wishes that Hurd could get some major funding from some source out there, or if it is going to continue as a pure hobbyist project, at least get a few more obsessive coders in there who can stick with it as long as is necessary.
(Score: 1, Funny) by Anonymous Coward on Friday April 17 2015, @05:43PM
Why is this a shame? Hurd is the Norma Desmond or Miss Havisham of the computer world. It holds its chin up regal and proud, but oblivious to the fact that time has past it by.
(Score: 2) by frojack on Friday April 17 2015, @08:02PM
Why is it disappointing?
(Look, I'm not trying to be a smart ass here, I really don't know about Hurd and I don't know a single person that ever even bothered to install it. What is it good for?)
No, you are mistaken. I've always had this sig.
(Score: 2, Funny) by Anonymous Coward on Saturday April 18 2015, @12:34AM
Like war, GNU Hurd is good for absolutely nothing.
(Score: 5, Insightful) by fnj on Friday April 17 2015, @10:27PM
Mach is a perfectly fine microkernel. The XNU kernel in every copy of OS X in the world incorporates Mach, and OS X works just fine.
The criticism of Mach performance dates back to a time when CPUs were vastly slower than X86s are now. I have a very strong idea that an OS based on straight Mach microkernel (not a hybrid like XNU) would not today be perceptibly slower in overall system performance than other OSs such as Windows, OS X, linux, or BSD. The clean, straightforward design of Mach is a large plus compared to microkernels such as L4, et al, incorporating elaborate tricks to speed up IPC - when I very much doubt that you can show on a current CPU such as a Haswell that IPC is any kind of a system performance problem at all.
The NeXT computer circa 1989 ran on a single core 25 MHz 68030; very advanced for its day yet ludicrously limited compared to even a multi-core GHz-level smart phone of today, let alone a Haswell/Broadwell laptop or desktop. The XNU kernel in NeXTSTEP was built on Mach 2.5 and nobody thought NeXTSTEP was slow even though the CPU had on the order of 1/1000 or less the performance of today's hardware.
Benchmarks were done decades ago. On an 486DX-50, a standard UNIX system call took an average of 21μs to complete, while the equivalent operation with Mach IPC averaged 114μs. But by simply scaling from 50 MHz to today's CPU speed of, say, 3 GHz suggests that the numbers become 0.35 and 1.9 microseconds respectively. Big deal. Now you're completely down in the noise.
(Score: 1, Interesting) by Anonymous Coward on Saturday April 18 2015, @12:46AM
The XNU kernel has essentially nothing to do with the original Mach microkernel. For example, the OSX kernel is monolithic.
I don't know why they just don't take Minix and just GPL it. That is the the only microkernel that is possibly going anywhere. It's developed by Tanenbaum and others who actually know what they are doing.
(Score: 2) by fnj on Saturday April 18 2015, @03:32AM
Wrong. XNU is a mix of micro and monolithic. Mach 3.0 is IN it, so it is ignorant to say it has nothing to do with "the original Mach microkernel".
(Score: 0) by Anonymous Coward on Saturday April 18 2015, @12:40PM
The NeXT era was 25 years ago! There has been a huge amount of code change since then. Just because the term "mach" is still used in some kernel-level function names it doesn't mean that much of the original Mach code is still present.
(Score: 2) by fnj on Saturday April 18 2015, @03:36PM
XNU in OS X is very much in the same form as XNU in NeXTSTEP. Both use Mach plus BSD kernel (and userland). The Mach part is upgraded from 2.5 to 3.0, and I/O Kit replaces Driver Kit.
In XNU, Mach code is used for:
thread and process preemption
protected memory
virtual memory management
inter-process communication
interrupt management
real-time support
kernel debugging support
console I/O
Those are pretty critical and far-reaching functions. So they chose not to route syscalls through Mach, and use the microkernel "server" model. That does not mean that Mach is not in there, and doing a lot of the things it was designed to do.
(Score: 2) by TheRaven on Saturday April 18 2015, @07:56PM
sudo mod me up
(Score: 2) by FatPhil on Saturday April 18 2015, @10:15AM
Sometimes it might be the modern languages and language-architecture libraries that are partly to blame, rather than the actual software architecture itself. I once straced a modern piece of bloatware which appeared to have locked up, and noticed that it was basically doing nothing but mmap/munmap. In a pretty-much-pessimal scatter-gather pattern. System calls constantly, tens of thousands per second. Add that all together and alas it's not noise.
Great minds discuss ideas; average minds discuss events; small minds discuss people; the smallest discuss themselves
(Score: 3, Interesting) by gnuman on Friday April 17 2015, @05:41PM
http://xkcd.com/1508/ [xkcd.com]
stylistic fixes" to the code-base
I'm sorry, but this shouldn't even make the changelong, never mind the news. Style fixes are the most benign of changes. What is the niche that HURD wants to occupy anyway?
(Score: 0) by Anonymous Coward on Saturday April 18 2015, @05:33AM
Oh, so it WILL be ready. 2060, hey? I can play with it in my dotage...
(Score: 0, Funny) by Anonymous Coward on Friday April 17 2015, @05:44PM
Polishing the turd.
(Score: 1) by pillo on Tuesday April 21 2015, @02:04PM
I hurd what you are doing there...
(Score: 2) by urza9814 on Friday April 17 2015, @06:11PM
...Why?
I don't mean that as flamebait or anything, I'm honestly curious. If I've got the history right, Hurd basically began because there were no open source Unix systems at the time. But we've got lots of options now, and Hurd sounds like it's about a decade behind at this point. With no signs that it's ever going to catch up. Is there some niche that Hurd fills or may fill, or at this point is it just something people work on solely because it's fun or interesting? Are there any use-cases which would really require a microkernel such that most mainstream *nix systems wouldn't work?
On an unrelated note...there's a car dealership around here named Hurd, and I always chuckle when I see their sticker on cars. Not really a name that would inspire much confidence to a geek! Of course, the other one I see all the time is "Tarbox", which seems even worse :)
(Score: 4, Funny) by CoolHand on Friday April 17 2015, @06:26PM
No systemd? :)
Anyone who is capable of getting themselves made President should on no account be allowed to do the job-Douglas Adams
(Score: 1, Redundant) by urza9814 on Friday April 17 2015, @06:37PM
I considered that, but there are still plenty of systemd-free Linux distros that seem to be far ahead of Hurd. For example, a two or three year old release of any distro that's been around that long :)
(Score: 2) by CoolHand on Friday April 17 2015, @07:19PM
Devuan is getting closer to a release.. I know they have some testing vm's to download and I think they have a testing .iso now and sound like it's getting close.. hopefully the succeed..
Anyone who is capable of getting themselves made President should on no account be allowed to do the job-Douglas Adams
(Score: 2) by present_arms on Friday April 17 2015, @07:24PM
No systemd here either :)
http://trinity.mypclinuxos.com/
(Score: 1, Informative) by Anonymous Coward on Friday April 17 2015, @07:48PM
You don't seem to be aware that, before they adopted (Linux-only) systemd, Debian had a spin that used a non-Linux kernel.
kFreeBSD [google.com]
Did that offer great utility relative to Torvalds' kernel?
You'll have to ask the folks that run/ran it.
-- gewg_
(Score: 2) by urza9814 on Monday April 20 2015, @07:05PM
Actually, I am aware of that -- used to have a couple servers running Debian kFreeBSD. But BSD is far ahead of HURD too, isn't it?
(Score: 3, Insightful) by Anonymous Coward on Friday April 17 2015, @06:49PM
Last time I read the Hurd webpage, it spoke mainly of how the developers are using it for research. There's nothing wrong with that; experimenting with a technically unneeded kernal which may or may not end up creating a useful product someday. At the very least it allows people to hone their skills.
(Score: 0) by Anonymous Coward on Friday April 17 2015, @06:11PM
Just like Linux 4.0 then.
(Score: 2, Insightful) by dime on Friday April 17 2015, @07:16PM
Live kernel patching infrastructure put into mainline is not a compelling new feature?
(Score: 0) by Anonymous Coward on Saturday April 18 2015, @12:36AM
Sounds like a huge security risk to me.
(Score: 2) by darkfeline on Saturday April 18 2015, @11:58PM
Computers are a huge security risk.
Actually, I take that back: humans are a huge security risk.
Seriously though, with good practices, I don't see live kernel patching as being a significant increase in risk. Even if it were, you can just compile Linux without it.
Join the SDF Public Access UNIX System today!
(Score: 0) by Anonymous Coward on Friday April 17 2015, @07:09PM
I'm not too sure where it leaves GNU Hurd but combining a modern service manager with a monolithic kernel offers the best* of both approaches. IPC and context switching overhead only where it makes sense - why bother with microkernels at all?
* "best" being a relative **cough** systemd **cough** term here
(Score: 2) by mtrycz on Friday April 17 2015, @09:12PM
Having a machine (or image) bootup in seconds (maybe even under a second, with modern hardware) is great for real-time scaling.
Your site is getting slashdotted? Add more workers in SECONDS. In REAL real-time.
Yes, we have the clouds now, and all that stuff, but it's an all round cool concept, and probably has still some innovations to offer. Hybrid kernels wouldn't have been possible without people pushing for micro, for example. I think there's still potential to be found there.
In capitalist America, ads view YOU!
(Score: 5, Insightful) by morgauxo on Friday April 17 2015, @08:10PM
I wonder how many OS kernels there are out there that somebody writes just because they want to or for the experience or something like that. Little lonely pieces of code sitting on their authors' hard drives never to be known to the outside world.
So with Hurd they still make it available if for some reason somebody does want to see it. What's wrong with that?
(Score: 0) by Anonymous Coward on Saturday April 18 2015, @12:48AM
Well, there's TempleOS. It's a ring-0 only operating system that the schizophrenic creator made to talk to God, but it's a genuinely interesting design.
(Score: 2) by kaszz on Saturday April 18 2015, @12:50AM
GNU Hurd is about exploring a concept that might turn out very useful. It might run circles around monolithic kernels in the future, and people would understand why if they understood the GNU Hurd concept. The code is irrelevant, the concept is important. When exploring new concepts one doesn't launch huge porting projects. One makes sure the concept is solid and then write from scratch on a suitable platform. Perhaps that will be ARM-64 or MIPS-64.
Anyone paying attention on the CPU front should realize how huge on-chip caches, multi-processing and GHz clocking have changed what type of systems approach is the best for optimal CPU usage. This is not DOS times where cache and more Hz is the only thing that matters. And where protected mode would impose a 3x penalty speed wise.
(Score: 2) by FatPhil on Saturday April 18 2015, @10:40AM
As a dabbler in the graphics demo scene in the early-mid 90s, all I can say is "whaaaaat!!??!?". Everyone used Tran's pmode, *everyone*.
Great minds discuss ideas; average minds discuss events; small minds discuss people; the smallest discuss themselves
(Score: 2) by kaszz on Saturday April 18 2015, @11:04AM
Have a look at the 80386 clock cycles per op-code for real vs protected mode.
(Score: 0) by Anonymous Coward on Saturday April 18 2015, @02:30AM
The thing about Linux is, it's huge and mostly opaque to most developers. There is room for an OS that's only a million LOC, sacrificing the kitchen sink for the ability to be molded by customers, but it has to be on x86_64 for starters not IA-32. And it better support LLVM and go as well as gcc.
(Score: 2) by mendax on Saturday April 18 2015, @04:48AM
Why do the GNU people even bother with Hurd. Yes, it's different. Yes, it's unusual. Yes, it may allow some very interesting things to occur. But, in an open-source, Unix-like OS world dominated by Linux and *BSD (let us not forget Darwin, the underpinnings of MacOS), and the amazing things that have been done with all of them, do we really need another Unix-like one?
Frankly, the most interesting and innovative operating system that has been developed in the last 25 years was BeOS. I don't see how Hurd could ever come close to it.
It's really quite a simple choice: Life, Death, or Los Angeles.
(Score: 3, Informative) by pTamok on Saturday April 18 2015, @08:38AM
You need to distinguish between the Operating System (OS) and the kernel.
Linux is the kernel, but the Operating System is mostly built on GNU utilities, which is why some people insist on referring to the combination as GNU/Linux.
Hurd is not an alternative Operating System, it is an alternative kernel. So you could have an operating system that uses Hurd as its kernel, and GNU utilities, and could therefore be referred to as GNU/Hurd.
The benefit of having a different kernel is that different kernels can be optimised for different things. Hurd is an example of a microkernel, which many people thought was a promising way of writing a 'better' (for limited values of better) kernel than 'monolithic' kernels. The Wikipedia article will tell you more about microkernels: http://en.wikipedia.org/wiki/Microkernel [wikipedia.org] - and indeed you can find out about monolithic kernels: http://en.wikipedia.org/wiki/Monolithic_kernel [wikipedia.org] and hybrid kernels: http://en.wikipedia.org/wiki/Hybrid_kernel [wikipedia.org] there too.