Debian proposes to drop 7 architectures

[n0cmonkey]

Originally posted by: Nothinman

In OpenBSD you can submit a formal bug for a port. Most people take the problems to the ports@ mailing list or the maintainer directly. I see two open PRs. OpenBSD takes ports seriously.

But AFAIK they're not supported, not audited and wouln't be considered release critical, right? Would a grave bug in a port delay a release?

It depends on what you mean by "supported." You can ask for help with a port on the mailing list like you could anything else.

They are not auditted the way that the source tree is, no. Some of the maintainers do their best going over them though.

I'm not sure if a port would delay a release. I do know there have been at least one broken port at release time (that was found out a bit too late to delay release) and Theo was not happy.

Huh?

Ignore thunderroller, he's a troll. He posts nonsensical posts in a lot of threads. Reporting him to the mods now.

 

[VirtualLarry]

Originally posted by: n0cmonkey
I still think it's a stupid idea.

Didn't GCC 3.x do something very similar, last year?
On one hand, I can understand that a lot of the support for these free-software projects is still volunteer, and given limits on that support, it should be spent most wisely, in the manner which will have the greatest overall benefit, but given how important that Debian is in the Linux world, being both one of the "most free" distributions, as well as the basis of perhaps the majority of "downstream" packaged/branded distributions based on Debian, this will certainly have a ripple effect.

I wonder how many embedded platforms run off of a variant of Debian too?

 

[n0cmonkey]

Originally posted by: VirtualLarry

Originally posted by: n0cmonkey
I still think it's a stupid idea.

Didn't GCC 3.x do something very similar, last year?

I think they dropped a couple of archs, and of course in GNU style they let some archs languish. I don't think they dropped 7 archs though, and not as many BIG ones as this proposal wants to do.

 

[R3MF]

it does appear debian was too broad to keep up the pace with stable releases, if cutting down the number of arch's helps this then more power to them.

 

[drag]

Originally posted by: Nothinman

Latest stable version of the linux kernel is 2.6.11.4 or something like that. The kernel in debian stable is what? 2.4.19?

2.6 packages are for sarge and sid available via apt if you want. And really kernel versions mean nothing to distributions, RHEL 3 was using 2.4.9 until very recently, RHEL 4 which was only released a few weeks ago is the first release from RH to even offer a 2.6 kernel.

That's the only way I can see Debian going with a timely release. Time-based releases are proven(OpenBSD model), feature-based releases are proven to be silly (Debian model).

OpenBSD doesn't support ports as much as Debian supports it's packages AFAIK. Time based release won't help them close RC bugs, there's currently nearly 700 RC bugs.

http://bugs.debian.org/release-critical/

Looks like thats only for overal packages. With 'Testing' it looks like they only have 127 bugs.

At least that's the way it appears to me.

I think that if they would of definately 'frozen' testing a few months ago they would of had these remaining bugs out of the way.

 

[Nothinman]

it does appear debian was too broad to keep up the pace with stable releases, if cutting down the number of arch's helps this then more power to them.

Number of supported arches has very little to do with the release delays, cutting them down will have virtually no affect.

 

[Nothinman]

Looks like thats only for overal packages. With 'Testing' it looks like they only have 127 bugs.

Ah, so it does.

 

[cleverhandle]

Originally posted by: Nothinman
Number of supported arches has very little to do with the release delays, cutting them down will have virtually no affect.

OK, so enlighten us. What is the cause of the release delays? And please say more than just "new d-i" - other distros have changed their installers, but they haven't gone close to three years between releases.

 

[Nothinman]

IMO the main problem is that most of the developers run sid, so there's no motivation for them to release. They always have what they need so getting sarge out the door falls behind on their priority list. Combine that with the fact that every time they decide "freeze time, release in 3 months" something like Gnome 2.8 comes out and everyone cries that Debian will get released with Gnome 2.6 so they push in 2.8 and restart the freeze.

There's no doubt that the number of architectures has an affect on the release, but it's not as big anyone would like think it is. 99% of the work is automated across the board by the buildd systems so the DDs never have to see the other arches unless there's a bug in their package.

And d-i is extremely relevant and was a nice big delay for sarge, it's probably the only tool that has to have completely different sections for each architecture. And with d-i they seperated everything, so the UI is essentially a plugin. The default will be ncurses I'm sure, but people were already working on other interfaces like GTK2. I'm not saying it should take 3 years to write an installer, but it's probably one of the biggest, work wise, projects in Debian.

And according to this DD, communication isn't as smooth as it could be: http://www.grep.be/blog/2005/02/01/#56903

 

[drag]

From what I see Nothingham is exactly right.

You see 90% of your applications your going to deal with are autotooled and will compile across many different platforms. Most of it is very automated..

this is the main reason that Debian still uses XFree86 and hasn't moved to X.org. It's because XFree86 is a monolythic thing.. no real automated way to install and test it. Many applications are dependant on it's libraries and it is non-modular. It makes it a pain in the @ss to upgrade it between versions. (luckly with X.org they are working towards a autotooled and modular free X server, which will make upgrading between versions much easier, and will benifit not just Debian but all distros)

And like it was stated before once you get a application that is written well enough to compile cleanly accross x86, x86-64, PowerPC, and Itanium it's pretty much garrenteed that it will work just fine on something like Mips or Arm.

(and if you think that stuff like ARM is very obscure, realise that XScale which is used in the majority of powerfull handheld devices is Arm archatecture)

One of the major benifits of Linux/Unix over something like Windows or VMS is that it's cross platform. When Windows has to move over to x86-64 it's a pain in the rear, but all this supporting dozens of different platforms is what enabled Linux to have a relatively stable AMD64 Suse distro the day that AMD started selling it's next gen platform to the public. Windows still hasn't got a non-beta AMD64 version yet, which was originally scedualed to be released 2 years ago.


What screws over Debian's release scedual is feature creep.

Just one more upgrade. Can we get KDE 3.4 included into testing, that would be realy cool!

That sort of thing.

If Debian would stick to it's guns about what is and what is not going to be included in the next stable release we wouldn't be discussing this issue.

The cross-platform thing is a red herring. It causes issues with the Debian Installer and such, but that is about it.

 

[Sunner]

I agree with dropping the really old arches, seriously who's gonna run anything halfway recent on an old 68040 or Motorolas old 88K stuff, same thing with s390, I wonder how many Debian installations are done on s390 boxes.

But kicking stuff like sparc64 out of there shouldn't happen IMO, Alpha, sparc64, and HP-PA are all still useful architectures that people actually do use for more than nostalgica purposes.

 

[drag]

... I wonder how many Debian installations are done on s390 boxes.

A few.

Not a whole lot, seeing how Mainframes are fairly sparse anyways. But they still exist.

 

[Sunner]

Originally posted by: drag

... I wonder how many Debian installations are done on s390 boxes.

A few.

Not a whole lot, seeing how Mainframes are fairly sparse anyways. But they still exist.

Yeah, buying an old US-II system on EBay is one thing, picking up a mainframe is another

I wonder if the number of Debian s/390 installations are few enough to count on the fingers on one hand, or if I'd need both hands

When questions like that can be asked without sarcasm, it's time to retire the port

 

[drag]

Originally posted by: Sunner

Originally posted by: drag

... I wonder how many Debian installations are done on s390 boxes.

A few.

Not a whole lot, seeing how Mainframes are fairly sparse anyways. But they still exist.

Yeah, buying an old US-II system on EBay is one thing, picking up a mainframe is another

I wonder if the number of Debian s/390 installations are few enough to count on the fingers on one hand, or if I'd need both hands

When questions like that can be asked without sarcasm, it's time to retire the port

I suppose so. But I am curious why they wanted to have it in the first place.

Looks like most of the work is taken up by a IBM guy, according to Debian's s390 port page. Looks like they just work at keeping a few Linux distros compatable with the s390 systems.

Also keep in mind that while you can get a old US-II system from E-bay, you can pick up a new Zseries from IBM. So at least they are still making computers for it!



That being said, I would be very suprised if S/390 would hold Sarge back any, even if it's a 'official' port.

 

[Sunner]

Yeah, IBM's mainframe division is actually seeing growth IIRC, though I sorta doubt a whole lot of those new zSeries boxes are shipped with Debian
SuSE/Redhat is probably on a bunch of them though.

 

[n0cmonkey]

I'd rather have m[68]8k support than s/390 support.

 

[rahvin]

Gentoo adds arch's, Debian drops them. Wonder which distribution is moving to the forefront.

 

[Nothinman]

I'd rather have m[68]8k support than s/390 support.

Depends on your outlook. m68k machines aren't being produced any more than I know of and AFAIK even right now they're barely usable for anything. And on the other side S/390 machines are still being produced and IBM is footing a large part of the development effort, they sort of have to since noone can afford the hardware. I'm not saying either should be dropped because as long as there are people willing to do the work why stop them? But if I had to choose one I would say m68k.

Gentoo adds arch's, Debian drops them. Wonder which distribution is moving to the forefront.

The day Gentoo moves to the forefront of Linux distributions is the day Linux gets set back a decade in progress.

 

[n0cmonkey]

Originally posted by: Nothinman

I'd rather have m[68]8k support than s/390 support.

Depends on your outlook. m68k machines aren't being produced any more than I know of and AFAIK even right now they're barely usable for anything. And on the other side S/390 machines are still being produced and IBM is footing a large part of the development effort, they sort of have to since noone can afford the hardware. I'm not saying either should be dropped because as long as there are people willing to do the work why stop them? But if I had to choose one I would say m68k.

Agreed with pretty much everything you said.

I can afford m68k machines (m88k I haven't researched lately, but those luna boxes are pretty), I can't afford an s/390.

 

[rahvin]

Gentoo adds arch's, Debian drops them. Wonder which distribution is moving to the forefront.

The day Gentoo moves to the forefront of Linux distributions is the day Linux gets set back a decade in progress.[/quote]

IMO it's nearly ahead of debian. Give it another year and it probably will be. Debian is so caught up in politics that they've forgotten about everything else.

 

[n0cmonkey]

Originally posted by: rahvin
IMO it's nearly ahead of debian. Give it another year and it probably will be. Debian is so caught up in politics that they've forgotten about everything else.

I don't think gentoo seems to have the strict love for F/OSS that Debian does. That's a problem, IMO.

 

[Nothinman]

IMO it's nearly ahead of debian. Give it another year and it probably will be. Debian is so caught up in politics that they've forgotten about everything else.

There's virtually no QA, there's half as many packages and AFAIK there's only binaries for big stuff like X and O, they don't care whether the included software is truly free or not, portage is missing some basic things that any package manager should be able to do and portage is dog slow compared to apt.