NetBSD 1.6 Has Been Branched

jschauma writes "Following Todd Vierlings announcement to the current-users Mailinglist, the NetBSD 1.6 Release Process has begun. This means not only that 1.6 has been branched off the cvs-tree, but also that daily snapshots will soon be available. Changes from 1.5 to 1.6 are listed here. A brief announcement including a best-case scenario release timetable is available from here. Whooot."

Excellent

[duffbeer703]

Dispite all of the blathering about how BSD is a dying breed of OS, the developers who have dedicated their time to BSD continue to make strides.

Hats off to the NetBSD team!

Re:*BSD Trolls are Dying!

[Yarn]

I've had a funny day, and that was still in the top 5 funny things that has happened, well done :)

Re:*BSD Trolls are Dying!

[Ded Bob]

For all practical purposes, Trolls are dead.

I wonder if a Turn Undead scroll would work. :)

Just a warning. The troll under the BSD bridge appears to dislike people: flea bite [slashdot.org]. I got de-karma-ized.

Re:Why *BSD is dying - an insider's view

[Strog]

Now we have to deal with OpenBSD trolls too?

Net and Free both use teams to control them. If someone gets pissed and leaves then life will go one for the rest. Several have left FreeBSD over the years and they keep moving forward.

If Theo decided that he was done or got fed up with all the other idiots out there(there's always idiots out there) then he could shut OpenBSD down tomorrow. Of course the source is all available and there are plenty of OpenBSD developers out there and maybe one or more would pick up the torch.

Theo still owns the copyright for the name and CD layout. The end result would be a fork of OpenBSD but how good would it be without Theo at the helm?

I love OpenBSD but *even* I realize that an OS is not a religion. They all have their respective areas for strengths and weaknesses. None are perfect and I think all 3 are very solid for their areas. I wouldn't pick OpenBSD for SMP or if I had a MIPS box but it sure would be a the top of the list for say bridging, firewall, routing, etc. and would be just fine for a single processor DHCP, webserver, AFS, etc.

Re:Why *BSD is dying - an insider's view

[Selmo]

OpenBSD has a focus. FeeBSD has lost its focus and is spread too thin over the map. One time they had a good OS for the ia32. But now that focus is lost. FreeBSD is way behind. NetBSD diverts too many resources to avenues of no return (thinking Atari, Amiga, 68K, and other marginal platforms). Long term, probably only OpenBSD will survive.

Did you know that NetBSD's ports collection [netbsd.org] is actually slightly larger than OpenBSD's? Everyone knows that FreeBSD is a cornucopia of packages....possibly even more than Debian [slashdot.org], depending on how you count.

Personally, I think if FreeBSD wants become more sucessful it needs to place the highest emphasis on porting to a few more architectures. Start with PPC and sparc64 since they're already fielded and widely used, then work on ia64 and x86-64 after that. I say this because x86 won't be around forever....look at all the consolidation and standardization among consumer PC's already...the trend is toward 'set top boxes' but I think it's been slowed with the tech economy's depression. Intel's insane price cuts on CPU's can only be hastening this devastation of profit margins.

With a commerical advocate in Wasabi Systems, NetBSD is poised to make a big splash in the growing embedded/wireless business and might well be the OS for anyone interested in wireless applications to focus on.

Mmmm, should be good

[LizardKing]

As I understand it, this will be the first release where all architectures are ELF based. No more recompiling the linker to avoid annoying warning messages on the Vax! The compiler will also be brought more upto date, and X will be version 4.2.0. Having followed the NetBSD security mailing list for a while, there will also be some nice little tweaks to the default install.

I'm a little bit unclear on whether this release will feature native threading support, which is the only API I'm missing from a certain other Unix-like operating system. Anyone reading know the score on threading support in 1.6?

Chris

Re: Native threading library ?

[Anonymous Coward]

No native threading yet. There is the nathanw_sa
branch to -current, which implements scheduler
activations to come up with a native thread
implementation. It's worth checking out
doc/BRANCHES (e.g. cvsweb:
http://cvsweb.netbsd.org/bsdweb.cgi/doc/B RANCHES?r ev=1.38&content-type=text/x-cvsweb-markup&only_wit h_tag=HEAD) and
read the pointers nathanw has put in that file
to understand the difference between this implementation,
what Linux initially chose, and user threading libs
like e.g. gnu pth.

IIrc, there was this question somewhen on current-users ML, too, and ppl have said
that it's likely there will be a new release (be it 1.7 or 2.0 or whatever) after quite short time
(compared to previous releases) which will merge
nathanw_sa and sommerfeld_i386_mp to both get
i386 MP support into it as well as the native
thread library.

Hth

any reason for slow releases

[Partisan01]

I follow BSD pretty regularly, and it seems like FreeBSD and OpenBSD churn out new releases every 6 months or so give or take, but from what i know of NetBSD it's been a while since they had a release. Does anyone know the reason for this, is it maybe due to less developers or integrating more things into the release. I remember reading on here a bit ago about zero buffering in the ethernet is it that stuff that's causing the hold up?

Re:any reason for slow releases

[LizardKing]

I don't know about FreeBSD, but OpenBSD follows a six monthly release cycle fairly rigidly, giving the developers a target for adding new features and stabilising them. It must also provide a crude way to estimate revenue, as CD sales obviously peak around the time of a release.

NetBSD has some different goals to the other BSD's. Correctness of the implementation of any new feature is valued highly, leading to a conservative development process. The SMP implementation is a good example - the NetBSD developers want to get it right, by comparison the Linux philosophy calls for early release of potentially buggy code. The Linux principle is that exposure to a wide audience will shake out bugs quickly. NetBSD relies on the fact that the most of the "audience" don't have the skill to fix bugs or even provide meaningful feedback. We could argue about which approach is best until the cows come home, but that's the way it is.

So in conclusion, NetBSD is released at a leisurely pace but this shouldn't be taken as a relection of how much development is going on.

Chris

Re:any reason for slow releases

[alan_d_post]

I think a better example is threads, where Linux has has a rather odd implementation, and NetBSD has no kernel support. When SA lands, it'll be good, and in the meantime . . . well . . . nobody is forcing you to use the system. :)

I think it's also good to point out that many people seem to do fine running -current, which releases as often as you want.