Stephan Kulow wrote:

On 27.08.2015 11:34, Andreas Hoffmann wrote:

...

Dear all,

does for the upcomming version 42.1 a 32 bit version will be available? Nor for milestone 1 I only can find a x64 version.

If 32 bit will be availalbe later on which stage of the development it is to be expected?

Beta? RC? I don't plan to create one at all

That's a pity. /Per -- Per Jessen, Zürich (26.2°C) http://www.dns24.ch/ - free dynamic DNS, made in Switzerland. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

And for me its the usage in small VMs, as 32 bit versions do need less RAM. Greetings Andreas On Thu, Aug 27, 2015 at 2:05 PM, Carlos E. R. wrote:

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256

On 2015-08-27 13:48, Per Jessen wrote:

...

Stephan Kulow wrote:

...

...

...

Beta? RC? I don't plan to create one at all

That's a pity.

Indeed. I have 32 bit servers.

- -- Cheers / Saludos,

Carlos E. R.

(from 13.1 x86_64 "Bottle" (Minas Tirith)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux)

iF4EAREIAAYFAlXe/QgACgkQja8UbcUWM1x8cAD/c2gRgjivOVunueZyjKxdwfZM 1x6hnO4Sx6fpMI2PSLUBAJ91FGnrYFr9yzxJYlF3GSEC/DyXsq5M1hgwmJF52UAO =WD76 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Same for me, I use 32 bits old laptops, because of price $ $ $ Dsant, from France On 08/27/2015 02:09 PM, Andreas Hoffmann wrote:

And for me its the usage in small VMs, as 32 bit versions do need less RAM.

Greetings Andreas

On Thu, Aug 27, 2015 at 2:05 PM, Carlos E. R. wrote:

...

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256

On 2015-08-27 13:48, Per Jessen wrote:

...

Stephan Kulow wrote:

...

...

Beta? RC? I don't plan to create one at all

That's a pity. Indeed. I have 32 bit servers.

- -- Cheers / Saludos,

Carlos E. R.

(from 13.1 x86_64 "Bottle" (Minas Tirith)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux)

iF4EAREIAAYFAlXe/QgACgkQja8UbcUWM1x8cAD/c2gRgjivOVunueZyjKxdwfZM 1x6hnO4Sx6fpMI2PSLUBAJ91FGnrYFr9yzxJYlF3GSEC/DyXsq5M1hgwmJF52UAO =WD76 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 08/27/2015 06:48 AM, Basil Chupin wrote:

On 27/08/15 22:05, Carlos E. R. wrote:

...

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256

On 2015-08-27 13:48, Per Jessen wrote:

...

Stephan Kulow wrote:

...

...

Beta? RC? I don't plan to create one at all

That's a pity. Indeed. I have 32 bit servers.

I really do not know what you were expecting. Afterall, you are simply a user :-) .

I learnt a new acronym today: DILLIGAF. The changes in KDE5 appear to be in this category - and it may be spreading.....

BC

Clever :^| What does the acronym stand for? -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

* Dsant [08-27-15 10:30]:

On 08/27/2015 04:27 PM, Carl Symons wrote:

...

On 08/27/2015 06:48 AM, Basil Chupin wrote:

...

On 27/08/15 22:05, Carlos E. R. wrote:

...

On 2015-08-27 13:48, Per Jessen wrote:

...

Stephan Kulow wrote:

...

>Beta? RC? I don't plan to create one at all

That's a pity. Indeed. I have 32 bit servers.

I really do not know what you were expecting. Afterall, you are simply a user :-) .

I learnt a new acronym today: DILLIGAF. The changes in KDE5 appear to be in this category - and it may be spreading.....

BC

Clever :^| What does the acronym stand for?

"It's an acronym for: Does It Look Like I Give a F@#k!!!!"

And is certainly uncalled for here. It is the type of post that *should* remain on the OPs computer and reflects *only* on his character. Shame. -- (paka)Patrick Shanahan Plainfield, Indiana, USA @ptilopteri http://en.opensuse.org openSUSE Community Member facebook/ptilopteri http://wahoo.no-ip.org Photo Album: http://wahoo.no-ip.org/gallery2 Registered Linux User #207535 @ http://linuxcounter.net -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Op donderdag 27 augustus 2015 23:48:40 schreef Basil Chupin:

I learnt a new acronym today: DILLIGAF. The changes in KDE5 appear to be in this category - and it may be spreading.....

BC

In my not so humble opinion you're crossing the lines of decent behaviour and respect. If you want to have the debate this way, it would have been polite to invite the KDE devs. An apology would be apropriate. My 2 cents -- Gertjan Lettink, a.k.a. Knurpht Official openSUSE Member openSUSE Forums Team -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Stephan Kulow [27.08.2015 12:27]:

On 27.08.2015 11:34, Andreas Hoffmann wrote:

...

Dear all,

does for the upcomming version 42.1 a 32 bit version will be available? Nor for milestone 1 I only can find a x64 version.

If 32 bit will be availalbe later on which stage of the development it is to be expected?

Beta? RC? I don't plan to create one at all

Greetings, Stephan

I see. Since SLE 12 is 64 bit only, it would be quite a lot of work to create a 32 bit branch for Leap. Werner --

32bit x86 , last officially released desktop/laptop cpu was released in 2010 , all 32bit consumer CPU are behind moral and physical service life . + No one report or repairs bug on 32bit openSUSE . On 28 August 2015 at 08:04, Felix Miata wrote:

Werner Flamme composed on 2015-08-28 06:48 (UTC+0200):

...

Stephan Kulow composed:

...

...

...

If 32 bit will be availalbe later on which stage of the development it is to be expected?

...

...

...

Beta? RC?

...

...

I don't plan to create one at all

...

I see. Since SLE 12 is 64 bit only, it would be quite a lot of work to create a 32 bit branch for Leap.

If I understand this, it means 32 bit users when 13.2 support and Evergreen support expire either switch to another distro, switch to the rolling release TW, or risk non-support? -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation)

Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!

Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Ondřej Súkup composed on 2015-08-28 08:25 (UTC+0200):

all 32bit consumer CPU are behind moral and physical service life .

Morals must be different in your country than in mine. There's no moral or physical reason here to junk a PC just because it's 5 years old. My newest of many functioning test PCs running openSUSE was manufactured ~6 years ago. The motherboard I'm typing this with is about 7-8 years old. When caps go bad, I replace the caps, not the motherboard. It's an eco-friendly way of life, maybe using more energy daily, but less energy in recycling processes, less consumption of scarce resources, and less energy and waste in landfill management. -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Hello, On Aug 28 09:07 Michal Kubecek wrote (excerpt):

... our limited resources ...

What exactly do you mean with "our"? I think: If you mean openSUSE contributors then those are free to spend their time on what they like to get done. In contrast if you mean SUSE employees it is out of scope when they spend SUSE working hours on issues that have no relevance for SUSE products. Of course SUSE employees are free to spend their leisure time on basically anything they like. Because SLE12 is only made for 64-bit, 32-bit specific issues have no relevance for SLE12 products and accordingly a 32-bit version of Leap as no relevance for SUSE. Regarding a 32-bit version of Leap made by openSUSE contributors see my other mail. This not any kind of official statement. It is only my personal point of view. Kind Regards Johannes Meixner -- SUSE LINUX GmbH - GF: Felix Imendoerffer, Jane Smithard, Graham Norton - HRB 21284 (AG Nuernberg) -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 On 2015-08-28 10:44, Michal Kubecek wrote:

And that I don't believe openSUSE Leap would suffer from an absence of i586 version.

Oh, yes, we would suffer. - -- Cheers / Saludos, Carlos E. R. (from 13.1 x86_64 "Bottle" (Minas Tirith)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) iF4EAREIAAYFAlXgN60ACgkQja8UbcUWM1x/aQD/adZt5X2izAB4B7k4fm50Woqm 8IMIul/jUZeMf6B2RssA+QGKSEZUX+K4B1FrVjSuGz+hskZLaywc/epLHyymUd8l =CmUT -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Dne 29.08.2015 v 6:06 Felix Miata napsal(a):

Michal Kubecek composed on 2015-08-28 10:44 (UTC+0200):

...

I don't believe openSUSE Leap would suffer from an absence of i586 version.

Goodwill and sacrifice for the benefit of fellow man do seem to have become concepts of historical importance.

It is very easy to sacrifice resources and time of others. M.

On Friday 28 of August 2015 10:15:03 Per Jessen wrote:

Michal Kubecek wrote:

...

And that's the point: most of those still running i586 distributions do run them on 64-bit capable hardware - because they believe

(a) they don't need x86_64 unless they have >4GB of memory (b) it consumes less memory (c) it consumes less disk space

It's not about belief, I can easily prove (b) to you. I don't care about (a) and (c).

OK, I'm pretty sure does consume _less_ memory. However, I seriously doubt you can prove to me the difference is (except for artificially crafted examples) big enough to outweigh the drawbacks of i586. Michal Kubeček -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Friday 28 of August 2015 11:11:37 Jan Matejka wrote:

On Fri, 28 Aug 2015 10:35:55 +0200 Michal Kubecek wrote:

...

OK, I'm pretty sure does consume _less_ memory. However, I seriously doubt you can prove to me the difference is (except for artificially crafted examples) big enough to outweigh the drawbacks of i586.

You can observe halved memory consumption on lightly loaded machines regardless of application.

No, I can't. Smaller, yes, but not halved, not by far. Michal Kubeček -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Friday 28 of August 2015 11:42:38 Jan Matejka wrote:

On Fri, 28 Aug 2015 11:25:11 +0200 Michal Kubecek wrote:

...

On Friday 28 of August 2015 11:11:37 Jan Matejka wrote:

...

You can observe halved memory consumption on lightly loaded machines regardless of application.

No, I can't. Smaller, yes, but not halved, not by far.

Yes. Halved. This is empiric data.

So are mine. When I last checked, the overall ratio was rather something like 1.25. I must admit this was ~5 years ago, I'm not even considering running an i586 system since I got rid of the last one (around 2008, IIRC). Michal Kubeček -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Michal Kubecek wrote:

On Friday 28 of August 2015 10:15:03 Per Jessen wrote:

...

Michal Kubecek wrote:

...

And that's the point: most of those still running i586 distributions do run them on 64-bit capable hardware - because they believe

(a) they don't need x86_64 unless they have >4GB of memory (b) it consumes less memory (c) it consumes less disk space

It's not about belief, I can easily prove (b) to you. I don't care about (a) and (c).

OK, I'm pretty sure does consume _less_ memory. However, I seriously doubt you can prove to me the difference is (except for artificially crafted examples) big enough to outweigh the drawbacks of i586.

Nothing artificial about it - in production, I can run many more postfix instances on 32bit than on 64bit. I have been doing this for more than ten years. In this environment, I haven't noticed any drawbacks of i586, but please do enlighten me. -- Per Jessen, Zürich (25.1°C) http://www.dns24.ch/ - your free DNS host, made in Switzerland. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

greg.freemyer@gmail.com wrote:

On August 28, 2015 5:53:33 AM EDT, Per Jessen wrote:

...

Michal Kubecek wrote:

...

...

Michal Kubecek wrote:

...

And that's the point: most of those still running i586 distributions do run them on 64-bit capable hardware - because

On Friday 28 of August 2015 10:15:03 Per Jessen wrote: they

...

...

believe

(a) they don't need x86_64 unless they have >4GB of memory (b) it consumes less memory (c) it consumes less disk space It's not about belief, I can easily prove (b) to you. I don't care about (a) and (c). OK, I'm pretty sure does consume _less_ memory. However, I seriously doubt you can prove to me the difference is (except for artificially crafted examples) big enough to outweigh the drawbacks of i586.

Nothing artificial about it - in production, I can run many more postfix instances on 32bit than on 64bit. I have been doing this for more than ten years. In this environment, I haven't noticed any drawbacks of i586, but please do enlighten me.

Per,

I'm very curious if a 64-bit kernel with primarily a 32-bit userspace would get you similar ram usage.

It seems reasonable for opensuse to create a hybrid 64-bit kernel / 32-bit userspace release.

From my perspective that seems like it would have very similar ram usage characteristics, but eliminate having the kernel having artificial memory constraints.

Hi Greg, I completely agree - it was suggested to me some time ago, and I have been wanting to try it out for a while. I think I did once have a look at building e.g. a 32bit postfix or apache on a 64bit system, I can't remember why I didn't pursue it. /Per -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Cristian Rodríguez wrote:

On Fri, Aug 28, 2015 at 6:53 AM, Per Jessen wrote:

...

Nothing artificial about it - in production, I can run many more postfix instances on 32bit than on 64bit. I have been doing this for more than ten years. In this environment, I haven't noticed any drawbacks of i586, but please do enlighten me.

So.. what's so special about postfix in this case ? there is an scalability problem.. that someone may be able to look at. This is not a reason to keep the i586 builds around, it is a reason to investigate why you can't do that in the 64bit version.

The 64bit version uses up more memory, that's all. -- Per Jessen, Zürich (22.5°C) http://www.hostsuisse.com/ - virtual servers, made in Switzerland. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Cristian Rodríguez wrote:

On Fri, Aug 28, 2015 at 5:30 PM, Per Jessen wrote:

...

Cristian Rodríguez wrote:

...

On Fri, Aug 28, 2015 at 6:53 AM, Per Jessen wrote:

...

Nothing artificial about it - in production, I can run many more postfix instances on 32bit than on 64bit. I have been doing this for more than ten years. In this environment, I haven't noticed any drawbacks of i586, but please do enlighten me.

So.. what's so special about postfix in this case ? there is an scalability problem.. that someone may be able to look at. This is not a reason to keep the i586 builds around, it is a reason to investigate why you can't do that in the 64bit version.

The 64bit version uses up more memory, that's all.

While indeed the size of pointer is bigger and that may/will have an impact in memory consumption.. Your problem looks very suspicious to me.. when it was the last time you tried to do this ?

(apologies for the delay in responding, it's been a very busy week). Last time - dunno, probably 3-4 years. I distinctly remember wanting to move up to 64bit away from PAE.

in which openSUSE version,

No idea. Maybe 11/12-something.

postfix version

Currently 2.11.x - back then whatever was the newest. Built from source.

and kernel version ?

Dunno.

ohh and what is the amount of postfix instances that the 32 bit machine can run vs the 64 bit one ? did you tried memory compression ?

I don't remember the numbers, something like 2000 on the 32bit machine vs 1200-1500 on the 64bit. Probably 4Gb or 8gb memory at the time. I'm really not sure, I didn't take any notes. -- Per Jessen, Zürich (16.8°C) http://www.dns24.ch/ - your free DNS host, made in Switzerland. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Cristian Rodríguez wrote:

On Fri, Sep 4, 2015 at 7:47 PM, Per Jessen wrote:

...

Last time - dunno, probably 3-4 years. I distinctly remember wanting to move up to 64bit away from PAE.

Ok..in 3 to 4 years both userspace and particularly the kernel change a huge deal and attempting to asses your problem with that frame of reference is going to be a futile endeavour, because we are effectively talking about something that no longer exists.

I haven't asked anyone to assess the problem. As to whether it exists, it's anybody's guess. I feel quite certain I can still run more postfix processes on a 32bit systems than on a 64bit with the same amount of memory. I haven't heard any real argument to the contrary. -- Per Jessen, Zürich (13.8°C) http://www.dns24.ch/ - free dynamic DNS, made in Switzerland. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

My 2 mythtv backend boxes are still running on Intel P4s. Very much alive and no moral nor physical reason to replace them. Ditto for our corporate firewalls and asterisk server.

try to calculate energy needs of intel P4 vs modern 64bit cpu with sufficent power for this backends .. On 28 August 2015 at 09:53, Per Jessen wrote:

Ondřej Súkup wrote:

...

32bit x86 , last officially released desktop/laptop cpu was released in 2010 , all 32bit consumer CPU are behind moral and physical service life .

My 2 mythtv backend boxes are still running on Intel P4s. Very much alive and no moral nor physical reason to replace them. Ditto for our corporate firewalls and asterisk server.

Sure, 32bit-only hardware will eventually die off, but I'd still like to run 32bit VMs and 32bit openSUSE on 64bit hardware.

...

+ No one report or repairs bug on 32bit openSUSE .

I report them when I come across them.

I think my most recent one was about 32bit xen install files missing in 13.2.

https://bugzilla.opensuse.org/show_bug.cgi?id=915963

-- Per Jessen, Zürich (22.4°C) http://www.hostsuisse.com/ - dedicated server rental in Switzerland.

-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Richard Brown wrote:

A few things I'd like to clear up

1) This is a decision by Coolo and other contributors involved in releasing openSUSE, not 'SUSE' 2) Therefore, if other contributors feel there is a need to continue making a 32-bit version of Leap, then they can do it 3) That said, Michal and others are making very good practical points why a 32-bit version makes little sense. So far, most of the arguments for keeping a 32-bit version seem to believe running 32-bit in some scenarios has greater benefits than they do in reality.

Anyone not using 32bit openSUSE in reality probably has little idea of the benefits.

number of users. Therefore if people do step up to continue making a 32-bit distribution, I'd like to either see them find a way of mitigating the hardware impact of their work, or help us find sponsors willing to provide more hardware for the Build Service and openQA -

My company has offered hardware including rack-space before, but apparently the only way is to contribute cycles is to load the stuff on a truck and haul it to Nürnberg. -- Per Jessen, Zürich (24.9°C) http://www.hostsuisse.com/ - dedicated server rental in Switzerland. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 On 2015-08-28 12:59, Takashi Iwai wrote:

IMO, we should soften the statement: e.g. i586 image may be provided, but only if anyone steps up as the image maintainer, and it'll be provided as a "Port", i.e. without any openQA and other tests performed for the official image.

It's important to show the attitude that we're willing to help, not only just throwing away.

Thanks. That would be some improvement :-) - -- Cheers / Saludos, Carlos E. R. (from 13.1 x86_64 "Bottle" (Minas Tirith)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) iF4EAREIAAYFAlXgSnoACgkQja8UbcUWM1z7lwD/deRE/W9onqcjmYTZHlYZq2tL PiIaWRDEYWFxnZG4WqoA+wY9ksuHbaUpBz83MS+17YA0dltkN0Mxf0pGCfJklOka =WhDE -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Fri, 2015-08-28 at 10:49 +0200, Richard Brown wrote:

2) Therefore, if other contributors feel there is a need to continue making a 32-bit version of Leap, then they can do it

Like the Raspberry Pi images. Its a fair point; and reasonable for the target of openSUSE - at developers and sysadmins, who will, most likely, be on the current hardware. Better for the official distro to target its target (and use its resources appropriately) So long as those same devs can use out tools (OBS, OpenQA, SUSE Studio) to build 32-bit (or any architecture) respins... I don't see an issue. -- James Mason Technical Architect, Public Cloud openSUSE Member SUSE jmason@suse.com ------------------------------------- SUSECon 2015: Register at susecon.com

Per Jessen writes:

James Mason wrote:

...

On Fri, 2015-08-28 at 10:49 +0200, Richard Brown wrote:

...

2) Therefore, if other contributors feel there is a need to continue making a 32-bit version of Leap, then they can do it

Like the Raspberry Pi images.

To my knowledge, we have separate ARM (and PowerPC) build-farms?

Not for armv6 which is built with qemu on x86-64 workers. Andreas. -- Andreas Schwab, SUSE Labs, schwab@suse.de GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE 1748 E4D4 88E3 0EEA B9D7 "And now for something completely different." -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Friday 28 of August 2015 02:04:31 Felix Miata wrote:

If I understand this, it means 32 bit users when 13.2 support and Evergreen support expire either switch to another distro

...or switch to the native architecture, finally.

switch to the rolling release TW...

I guess the only thing that really surprises me is that dropping i586 hasn't been done in Tumbleweed first.

or risk non-support?

As I wrote already some time ago, I'm not very confident about the level of i586 openSUSE support we have been providing for the last few years. Sure, we may run it through openQA (actually, I'm not even sure about that) but how many beta testers run it on their machines (compared to x86_64)? How likely are you going to get help with an i586-specific bug? I, for one, would have to install such system first as I haven't been running one for years. And I wouldn't be too happy about it as 32-bit address space is severely limiting which requires a lot of ugly hacks, hindering the debugging severely. I'm afraid discontinuing i586 would be something I would call acknowledging the state of things and stopping pretending rather than some big and groundbreaking step. Michal Kubeček -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Friday 28 of August 2015 03:16:41 Felix Miata wrote:

Michal Kubecek composed on 2015-08-28 08:48 (UTC+0200):

...

I'm afraid discontinuing i586 would be something I would call acknowledging the state of things and stopping pretending rather than some big and groundbreaking step.

Last I checked every distrowatch listing above openSUSE still provided a 32 bit version, and IIRC, it was necessary to drop below the top 10 to find one that didn't. Fedora still calls its 386.

(Question of distrowatch listing order relevance aside.) Well, this doesn't actually contradict what I said. I have little reason to believe their level of actual support is much different from ours.

I remember seeing plenty of threads around where this sort of topic has come up, and seeing lots of people still using 32 bit by preference on their 64 bit systems.

That's exactly what I suspected: most people using 32-bit distributions these days actually do so because of their beliefs, not because they have to. Such use is certainly legitimate - but way less relevant for the question whether we should support the architecture.

It seems to me any acknowledgement would primarily be that the majority of developers don't want to bother with less than the newest and fastest machines,

Newest and fastest? 64-bit CPU's are widely available since ~2003 and prevailing since ~2005, for last 5 years, it's almost impossible to buy a 32-bit one. And again, it's not about "fastest". The tricks kernel has to do to cope with 32-bit architecture are quite ugly. There are even problems that can't be resolved on i586 (I remember a guy having over 60% of his 2GB RAM unused but unable to add a netfilter rule because of memory allocation failure). You have fewer registers, leading to much less efficient function calling convention etc.

largely I'll bet to compensate for software bloat, driving the cycle that makes vendors happy, but not so much users, particularly those on tight budgets.

Seriously?

It won't surprise me if whichever top 10 distro first dumps 32 bit falls at least 2 spots in short order, unless several do it in short order.

Pure speculation (or even wishful thinking). I could say once someone dares to do it, several others will follow - and it would be about as much fact based as your claim (i.e. not at all). Michal Kubeček -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Friday 28 of August 2015 12:48:58 Carlos E. R. wrote:

On 2015-08-28 10:30, Michal Kubecek wrote:

...

And again, it's not about "fastest". The tricks kernel has to do to cope with 32-bit architecture are quite ugly. There are even problems that can't be resolved on i586 (I remember a guy having over 60% of his 2GB RAM unused but unable to add a netfilter rule because of memory allocation failure). You have fewer registers, leading to much less efficient function calling convention etc.

Most or all of the 32 bit hardware have less than 4 GB RAM, so you could drop the hacks needed to support more RAM. It would not impact me, I assume.

I wasn't talking about PAE, that would be yet another layer of ugliness. Michal Kubeček -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512 On Fri, 28 Aug 2015 13:50:48 +0200 "Carlos E. R." wrote:

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256

On 2015-08-28 12:55, Michal Kubecek wrote:

...

On Friday 28 of August 2015 12:48:58 Carlos E. R. wrote:

...

On 2015-08-28 10:30, Michal Kubecek wrote:

...

And again, it's not about "fastest". The tricks kernel has to do to cope with 32-bit architecture are quite ugly. There are even problems that can't be resolved on i586 (I remember a guy having over 60% of his 2GB RAM unused but unable to add a netfilter rule because of memory allocation failure). You have fewer registers, leading to much less efficient function calling convention etc.

Most or all of the 32 bit hardware have less than 4 GB RAM, so you could drop the hacks needed to support more RAM. It would not impact me, I assume.

I wasn't talking about PAE, that would be yet another layer of ugliness.

Doesn't the compiler takes care automatically of having less registers? :-o

Yes. But you still need to test the kernel and compilers. Then still some corner cases may fall through which affect only some applications triggering the compiler/kernel bug. Then some applications may inline assembly in the code, which would be another thing to look for.

I don't suppose a package maintainer has to do anything there if needed: it would be an upstream task.

In theory yes, in practice the package maintainers have to flip their own burgers sometimes. - -- Jan Matějka | QA Engineer for Maintenance SUSE LINUX s.r.o. | https://www.suse.com/ GPG: A33E F5BC A9F6 DAFD 2021 6FB6 3EBF D45B EEB6 CA8B -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iQEcBAEBCgAGBQJV4FitAAoJEIN+7RD5ejah2bMH/13nFhIataKk0w7dqGZUtwxO AJeeqhAbZS9vciGrqg4LO/vddoHrby27CYAMOzm4HGttFu2EoDM8YxckS4GBv0se 2/c4O3bM3cLYYMa2tJQuIIgZXig8M/tMe+N4pnH/csBQmamBjs+aMDaatmrdXR60 F81IWEbUIrL/PTuHLTqQcs9IuVucqt5Mw1j3I8sad3dWt0uhDWIqp5koKKVB6RCr 9XY1VTWD4f+vHIPFU/3hNi25tbhGoGUh+zVA7apwNld/5x6/LMg6HdixS76oiLT4 pznky7JpCpXL1D2hV7PuiMvic0s8oo9in7wgZ5wBlkkzw2xtODXeF7uQ6poIjHY= =jCE0 -----END PGP SIGNATURE----- N�����r��y隊Z)z{.���r�+�맲��r��z�^�ˬz��N�(�֜��^� ޭ隊Z)z{.���r�+��0�����Ǩ�

On 08/28/2015 10:30 AM, Michal Kubecek wrote:

That's exactly what I suspected: most people using 32-bit distributions these days actually do so because of their beliefs, not because they have to.

No, because I buy old 32 bits hardware... Dsant, from France -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 28 August 2015 at 13:01, Dsant wrote:

No, because I buy old 32 bits hardware...

Dsant, from France

I'm confused where the expectation comes from that new modern operating systems should run on old obsolete hardware 32 bit hardware is no longer produced - this is the very definition of obsolete. Now we have clear indications that the use base of 32-bit Linux is reaching minimal levels, I really do not see the justification for the extra work that supporting the 32-bit intel architecture requires -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Fri, 28 Aug 2015 15:35:10 +0200, Richard Brown wrote:

On 28 August 2015 at 13:01, Dsant wrote:

...

No, because I buy old 32 bits hardware...

Dsant, from France

I'm confused where the expectation comes from that new modern operating systems should run on old obsolete hardware

Well, it's been a virtue of Linux in general that it keeps supporting the old hardware the other OSes discontinued. Takashi -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 08/28/2015 03:35 PM, Richard Brown wrote:

Now we have clear indications that the use base of 32-bit Linux is reaching minimal levels, I really do not see the justification for the extra work that supporting the 32-bit intel architecture requires Ok, I believe you, you convinced me because you might have figures.

I'm confused where the expectation comes from that new modern operating systems should run on old obsolete hardware

32 bit hardware is no longer produced - this is the very definition of obsolete. Explanation : the second hand laptops market, only about 20€ per machine. Anyway it will happens one day, so 32 bits is not worth.

Dsant, from France -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 28.08.2015 15:35, Richard Brown wrote:

On 28 August 2015 at 13:01, Dsant wrote:

...

No, because I buy old 32 bits hardware...

Dsant, from France I'm confused where the expectation comes from that new modern operating systems should run on old obsolete hardware

32 bit hardware is no longer produced - this is the very definition of obsolete.

While it may not see wider use, there is x86 hardware produced which is 32-bit [1]. You can easily run new distributions on it and play with such systems.

Now we have clear indications that the use base of 32-bit Linux is reaching minimal levels, I really do not see the justification for the extra work that supporting the 32-bit intel architecture requires

[1] http://ark.intel.com/de/products/79084/Intel-Quark-SoC-X1000-16K-Cache-400-M... -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 28.08.2015 17:23, Ondřej Súkup wrote:

ad [1] http://ark.intel.com/de/products/79084/Intel-Quark-SoC-X1000-16K-Cache-400-M...

is SoC for microcontrollers and with BUGGY instructions for threading , not usable in any normal distribution -> https://en.wikipedia.org/wiki/Intel_Quark#Segfault_bug

Compile glibc with -march=i486 which is indeed not a desired behavior, but with this done Opensuse 13.2 works fine on the Quark X1000 (Intel Galileo Board)! This could be done as a replacement for the normal glibc, to provide full featured distributions for such systems.

+ Bonus , all AMD64 cpus support 32bit x86 but nobody want support three architectures in one, 32bit , 64bit and x32 .. i too expensive without real benefits

On 28 August 2015 at 16:44, Tobias Klausmann wrote:

...

On 28.08.2015 15:35, Richard Brown wrote:

...

On 28 August 2015 at 13:01, Dsant wrote:

...

No, because I buy old 32 bits hardware...

Dsant, from France I'm confused where the expectation comes from that new modern operating systems should run on old obsolete hardware

32 bit hardware is no longer produced - this is the very definition of obsolete.

While it may not see wider use, there is x86 hardware produced which is 32-bit [1]. You can easily run new distributions on it and play with such systems.

...

Now we have clear indications that the use base of 32-bit Linux is reaching minimal levels, I really do not see the justification for the extra work that supporting the 32-bit intel architecture requires

[1] http://ark.intel.com/de/products/79084/Intel-Quark-SoC-X1000-16K-Cache-400-M...

-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 07.09.2015 23:52, Andreas Färber wrote:

Hi Tobias,

Am 28.08.2015 um 17:53 schrieb Tobias Klausmann:

...

On 28.08.2015 17:23, Ondřej Súkup wrote:

...

ad [1] http://ark.intel.com/de/products/79084/Intel-Quark-SoC-X1000-16K-Cache-400-M...

is SoC for microcontrollers and with BUGGY instructions for threading , not usable in any normal distribution -> https://en.wikipedia.org/wiki/Intel_Quark#Segfault_bug Compile glibc with -march=i486 which is indeed not a desired behavior, but with this done Opensuse 13.2 works fine on the Quark X1000 (Intel Galileo Board)! This could be done as a replacement for the normal glibc, to provide full featured distributions for such systems. Do you actually have openSUSE running on a Galileo board, or is this just theory? I tried to get i586 openSUSE running as a Hackweek project [1] and never succeeded.

Which grub, which kernel, ...?

Thanks, Andreas

[1] https://hackweek.suse.com/12/projects/928

Hi Andreas, no this was not just theory, i have a repo which has, maybe had a working 13.2 image ready [1] (havent looked for some time if the image still works) for a Galileo Gen 1. Things done: - ported most parts of the kernel patches from the Yocto Linux to the latest stable kernel (3.18.5 at that time), named kernel-galileo. The patches are not really clean, but as of now, this shouldnt matter as newer kernels should work with the Galileo anyway! - Adjusted glibc with the -march=i486 (i am looking for a better version though). - Modify kiwi to build smaller appliances especially for the Galileo - Create a tiny appliance which boots on a Galileo machine with several different sizes predefined (4, 8, 16, 32GB root fs size) [2]. The outcome was a small appliance with software for a first boot, after that you could easily do a zypper in or zypper up (glibc had to be locked though). [1] https://build.opensuse.org/project/monitor/home:tobijk:Galileo:13.2 [2] http://download.opensuse.org/repositories/home:/tobijk:/Galileo:/13.2/images... -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Friday 28 of August 2015 10:28:11 Larry Finger wrote:

As much of the world cannot afford the latest shiny boxes, that will take a while.

People, please... Some of these "latest shiny boxes" you are talking about are over 10 years old. They do not shine at all and a lot of them has been discarded since years ago. OK, I don't deny there are still people using 32-bit hardware (even if I'm very sceptical about number of those) but calling 64-bit machines in general "latest shiny boxes"? Seriously? Michal Kubeček -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Michal Kubecek composed on 2015-08-28 17:38 (UTC+0200):

Larry Finger wrote:

...

As much of the world cannot afford the latest shiny boxes, that will take a while.

People, please... Some of these "latest shiny boxes" you are talking about are over 10 years old. They do not shine at all and a lot of them has been discarded since years ago. OK, I don't deny there are still people using 32-bit hardware (even if I'm very sceptical about number of those) but calling 64-bit machines in general "latest shiny boxes"? Seriously?

Not everyone who needs a PC gets to buy or even choose one. If it's new to them, it's new, and maybe shiny as well. What they can much more likely choose is the FOSS they put on what they do have. Whichever major distro is the first to drop 32 bit becomes to those people the distro for the elite. Those that fall freely into the hands of such people likely got some sort of refurbing at the hands of people keeping things simple as practical, likely using the same arch for everything crossing their paths. -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Mathias Homann composed on 2015-08-29 07:35 (UTC+0200):

Felix Miata composed on 2015-08-28 23:49 (UTC-0400):

...

Not everyone who needs a PC gets to buy or even choose one. If it's new to them, it's new, and maybe shiny as well. What they can much more likely choose is the FOSS they put on what they do have. Whichever major distro is the first to drop 32 bit becomes to those people the distro for the elite.

... Red Hat Enterprise Linux has dropped x86 32bit support with the release of RHEL7

When I use the word "distro", it's referring to FOSS. SLE and RHEL are not FOSS as I understand its meaning. -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 On 2015-08-29 08:44, Cristian Rodríguez wrote:

On Sat, Aug 29, 2015 at 3:23 AM, Felix Miata wrote:

...

When I use the word "distro", it's referring to FOSS. SLE and RHEL are not FOSS as I understand its meaning.

they *are* FOSS for are practical purposes.

You can not install them in refurbished machines, those given to people with scarce resources, or in the third world. Or is SUSE giving OS donations? - -- Cheers / Saludos, Carlos E. R. (from 13.1 x86_64 "Bottle" (Minas Tirith)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) iF4EAREIAAYFAlXhpjgACgkQja8UbcUWM1xZiAD/bYcj+FeGdbmsfEdmXs9Nc1T4 qNVlfTj8p1ez4AgLPTkBAIgBNgSdyq6C9qpqI21pSbfeLPkq082KThy/u3EieE3K =772k -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 On 2015-08-29 18:07, Michal Kubecek wrote:

On Sat, Aug 29, 2015 at 02:31:52PM +0200, Carlos E. R. wrote:

...

You can not install them in refurbished machines, those given to people with scarce resources, or in the third world. Or is SUSE giving OS donations?

SUSE is not selling SLE licenses, you can download it and install it anywhere you wish. What are customers paying for are support subscriptions, not software licenses.

I know that. It doesn't change the meaning of what I said: you can not install SLES/SLED in donated machines for people with scarce resources, because they can't keep them updated. - -- Cheers / Saludos, Carlos E. R. (from 13.1 x86_64 "Bottle" (Minas Tirith)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) iF4EAREIAAYFAlXh/dsACgkQja8UbcUWM1xOrgD/aTPNCUn8iGZq8MmN2sXh9JG8 yVadH8W+kYSrXSMUcsgA/jG9w3QcuU8+6D+CEy1DSkOBlAcueguqbi99mpx6/Bg7 =Y1nQ -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Am 28.08.15 um 17:38 schrieb Michal Kubecek:

People, please... Some of these "latest shiny boxes" you are talking about are over 10 years old. well long discussion about it, still no result.

But the truth is, that a ThinkPad T60 or X60s with Core Duo and 4 GB RAM plus SSD still is a very good system for doing daily work. After Windows XP was discontinued, I mirgrated 6 or 7 systems lilke these to openSUSE (32 bit) and all people using them are happy with what they have now. It's just doinig what it should and I mostly have no work with it. Unfortunately only for 18 month, then I have to do a "major" system upgrade was all possible problems. In my office we are using CentOS because of that, but personally I like openSUSE much more. So Leap would be THE solution and I'm waiting for it to install it (at least RC1) and, a 32 bit version also would be _very_ welcome (CentOS 7 also will be available as 32 bit version)! So just my 2 cent ... -- kind regards, Thorolf -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Tue, 2015-10-13 at 14:50 +0200, Carlos E. R. wrote:

On 2015-10-13 10:52, Thorolf Godawa wrote:

...

But the truth is, that a ThinkPad T60 or X60s with Core Duo and 4 GB RAM plus SSD still is a very good system for doing daily work.

That's a 32 bit cpu from 2006-2008.

No, the whole Intel Core microarchitecture is 64Bits. I'm running 64Bit TW on a Core2Duo MacBook (Penryn T8xxx) just fine.

-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Tue, 2015-10-13 at 15:31 +0200, Richard Brown wrote:

On 13 October 2015 at 15:26, Johannes Thumshirn wrote:

...

On Tue, 2015-10-13 at 14:50 +0200, Carlos E. R. wrote:

...

On 2015-10-13 10:52, Thorolf Godawa wrote:

...

But the truth is, that a ThinkPad T60 or X60s with Core Duo and 4 GB RAM plus SSD still is a very good system for doing daily work.

That's a 32 bit cpu from 2006-2008.

No, the whole Intel Core microarchitecture is 64Bits. I'm running 64Bit TW on a Core2Duo MacBook (Penryn T8xxx) just fine.

Core Duo (Yonah) = 32-bit Core 2 Duo = 64-bit

Ah sorrt, I misread. My fault

Source: wikipedia https://en.wikipedia.org/wiki/Yonah_(microprocessor)

-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Tue, 13 Oct 2015 15:02, Richard Brown wrote:

On 13 October 2015 at 10:52, Thorolf Godawa wrote:

...

Am 28.08.15 um 17:38 schrieb Michal Kubecek:

...

People, please... Some of these "latest shiny boxes" you are talking about are over 10 years old. ...

CentOS 7 is only available as a 64-bit version

https://wiki.centos.org/Download

I hear rumours of a 'community created' 32-bit version, but like this thread has spelled out several times now, we're open to that idea at openSUSE too, just none of those people previously involved in making our releases is interested in it..

Announcement today: C7 for 32bit available as AltArch: Wiki: (there are some Bugs, please read beforehand) https://wiki.centos.org/SpecialInterestGroup/AltArch/i386 DL-ISO: http://mirror.centos.org/altarch/7/isos/i386/ LEAP 42.1 will get some AltArchs sooner or later: AArch64, arm7, ppc64, etc. IMHO there is the place where the x86_32 arch should reside. - Yamaban. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Am 13.10.15 um 15:02 schrieb Richard Brown:

CentOS 7 is only available as a 64-bit version

https://wiki.centos.org/Download well, it was available as beta for some time, but because I use CentOS exclusively on servers, the 32-bit version of CentOS 7 is not important for me.

Beginning with CentOS 6 I installed all servers with the 64-bit version and for servers 64-bit are mandatory and (really) old hardware is nothing you have to take care in business use. But as I mentioned before, for personal use an old laptop or PC, even with a 32-bit CPU is an suitable option! And for these systems, openSUSE is one of the best options to use, preferable as LTE version.

I hear rumours of a 'community created' 32-bit version, but like this thread has spelled out several times now, we're open to that idea at openSUSE too, just none of those people previously involved in making our releases is interested in it.. Just got this:

Betreff: [CentOS-announce] CentOS Linux 7 for 32-bit x86 (i386) Architecture Datum: Mon, 12 Oct 2015 11:07:17 -0500 Von: Johnny Hughes Antwort an: centos@centos.org Organisation: The CentOS Project An: CentOS-Announce We would like to announce the general availability of CentOS Linux 7 for the 32-bit x86 (i386) architecture. This is the first major release of the 32 bit x86 by the AltArch Special Interest Group. This release is based on the Source Code from the CentOS 7 (1503) x86_64 architecture and includes all current updates from the main CentOS 7 tree. Installation ============ Install ISOs are available here: http://mirror.centos.org/altarch/7/isos/i386/ sha256sums: CentOS-7-i386-DVD-1503.iso: 8973b31b45878b2277c19465b0260aab4efb86126033ee0909fafa47ace5c3af CentOS-7-i386-Everything-1503.iso: 1c016cd95aaa8fe9bc68376f388c03c704eacd3f303e7f37f026e6aed4f78b77 CentOS-7-i386-Minimal-1503.iso: 4b71900eda2e863234e2edee9b2c3753f83bc9a0912551d26f69ff1e202ca979 CentOS-7-i386-NetInstall-1503.iso: cea236c94a26e62330acbbf8fe0e1aefa75f282fdbaab7f169213e8fb33d2cc4 The install process is identical to CentOS 7 x86_64 via kickstart or ISO installers. ... -- kind regards, Thorolf -- Chau y hasta luego, Thorolf -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 2015-08-28 13:01, Dsant wrote:

On 08/28/2015 10:30 AM, Michal Kubecek wrote:

...

That's exactly what I suspected: most people using 32-bit distributions these days actually do so because of their beliefs, not because they have to.

No, because I buy old 32 bits hardware...

Are you also paying old utility bills? :) Of course, it depends on what percentage of time the machine is running and it varies among countries, but you might want to include the cost of the electricity consumed in your buying decisions. Besides, old x86_64 hardware has been available for some time now as well. Michal -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Michal Kubecek composed on 2015-08-28 10:30 (UTC+0200):

Felix Miata wrote:

...

I remember seeing plenty of threads around where this sort of topic has come up, and seeing lots of people still using 32 bit by preference on their 64 bit systems.

That's exactly what I suspected: most people using 32-bit distributions these days actually do so because of their beliefs, not because they have to. Such use is certainly legitimate - but way less relevant for the question whether we should support the architecture.

Maybe there are more "have tos" than you think. Did upgrading an existing 32 bit installation to 64 bit ever become a supported option? I don't recall any such thing. I find upgrades far less painful than fresh installations. Probably other 32 bit preferrers do too.

...

It seems to me any acknowledgement would primarily be that the majority of developers don't want to bother with less than the newest and fastest machines,

Newest and fastest? 64-bit CPU's are widely available since ~2003 and prevailing since ~2005, for last 5 years, it's almost impossible to buy a 32-bit one.

The point was that devs who might otherwise wish to at least test on older systems find plenty reasons not to spend the time, among which waiting on slower RAM, CPU and I/O busses, not to mention logistics of keeping older equipment functional and pragmatically available.

And again, it's not about "fastest". The tricks kernel has to do to cope with 32-bit architecture are quite ugly. There are even problems that can't be resolved on i586 (I remember a guy having over 60% of his 2GB RAM unused but unable to add a netfilter rule because of memory allocation failure). You have fewer registers, leading to much less efficient function calling convention etc.

...

largely I'll bet to compensate for software bloat, driving the cycle that makes vendors happy, but not so much users, particularly those on tight budgets.

Seriously?

I never asked for compositing, the bling it enables, or the speed and RAM they depend on. KDE3 and TDE still match my needs as well as ever, unlike KDE4 and Plasma5. Did the kernel really need to more than double in size in 8 years[1], plucking so much out of userspace? Or was that growth a consequence of software bloat needing compensatory assistance to keep newer, faster PCs from seeming slower? There do still exist people who don't like having fixed what ain't broke. :-) [1] https://en.wikipedia.org/wiki/Linux_kernel#/media/File:Lines_of_Code_Linux_K... -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Sat, Aug 29, 2015 at 05:52:20PM +0200, Christian Boltz wrote:

Hello,

Am Samstag, 29. August 2015 schrieb Felix Miata:

...

Maybe there are more "have tos" than you think. Did upgrading an existing 32 bit installation to 64 bit ever become a supported option? I don't recall any such thing.

It isn't officially supported, but I know that some people (including me) did it successfully, so it should work ;-)

I suppose if we don't release an official i586 Leap image, we should consider testing and supporting the i586 13.2 -> x86_64 Leap migration, either direct or two-step one (via x86_64 13.2). Michal Kubeček -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Fri, 28 Aug 2015 03:16:41 -0400, Felix Miata wrote:

Last I checked every distrowatch listing above openSUSE still provided a 32 bit version, and IIRC, it was necessary to drop below the top 10 to find one that didn't.

Distrowatch is not a measure of distribution popularity or userbase. Distrowatch is a page counter that measures the popularity of pages for particular distributions on distrowatch. That's all they measure. People seem to think that there's some magic about Distrowatch rankings that has something to do with how many people run a particular distribution. That's entirely false.

From the distrowatch "Page Hit Ranking" page:

"The DistroWatch Page Hit Ranking statistics are a light-hearted way of measuring the popularity of Linux distributions and other free operating systems among the visitors of this website. They correlate neither to usage nor to quality and should not be used to measure the market share of distributions. They simply show the number of times a distribution page on DistroWatch.com was accessed each day, nothing more." I get really tired of people confusing Distrowatch's statistics with distribution popularity or other such metrics. If I visit the Fedora page, it counts a hit on the Fedora page. I'm not using Fedora, but I might have a question about what the latest version is or what's included in it, and distrowatch can tell me that. Since I don't use Fedora at all, it's an incorrect assumption to draw that the increase I give the Fedora page by visiting that page has anything to do with the number of people using Fedora. Interest in a distribution is not the same as using it. That is obvious. Understanding that Distrowatch at best measures *interest* and not *usage* seems to be far less obvious to a lot of people. Jim -- Jim Henderson Please keep on-topic replies on the list so everyone benefits -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Michal Kubecek writes:

On Fri, Aug 28, 2015 at 03:16:41AM -0400, Felix Miata wrote:

...

Michal Kubecek composed on 2015-08-28 08:48 (UTC+0200):

...

I'm afraid discontinuing i586 would be something I would call acknowledging the state of things and stopping pretending rather than some big and groundbreaking step.

Last I checked every distrowatch listing above openSUSE still provided a 32 bit version, and IIRC, it was necessary to drop below the top 10 to find one that didn't. Fedora still calls its 386.

They may call it so but as 386 CPU support in mainline kernel was dropped in December 2012, I seriously doubt they make the effort to keep maintaining all the hacks needed to actually run on 386.

They actually require 686+. Andreas. -- Andreas Schwab, SUSE Labs, schwab@suse.de GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE 1748 E4D4 88E3 0EEA B9D7 "And now for something completely different." -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Michal Kubecek wrote:

On Friday 28 of August 2015 02:04:31 Felix Miata wrote:

...

If I understand this, it means 32 bit users when 13.2 support and Evergreen support expire either switch to another distro

...or switch to the native architecture, finally.

...

switch to the rolling release TW...

I guess the only thing that really surprises me is that dropping i586 hasn't been done in Tumbleweed first.

...

or risk non-support?

As I wrote already some time ago, I'm not very confident about the level of i586 openSUSE support we have been providing for the last few years. Sure, we may run it through openQA (actually, I'm not even sure about that) but how many beta testers run it on their machines (compared to x86_64)?

If the level of testing and/or QA is an indicator for whether we should include <something>, there are a lot things that need to left out.

How likely are you going to get help with an i586-specific bug?

Well, how likely am I to hit an i586-specific bug? Also, if the probability of getting timely help with an xxx-specific bug is an indicator, many things should be left out. IMHO, neither is an applicable argument. -- Per Jessen, Zürich (22.4°C) http://www.hostsuisse.com/ - dedicated server rental in Switzerland. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

I wonder what might happen for things that are cross-compiled on OBS. Like all the MinGW stuff. I would imagine that there is no reason that the 32-bit compiles of the Windows versions of things could not continue on 64-bit hosts. We actually use those things. I do not know that all our users have 64-bit Windows installed. Roger Oberholtzer RST Systems Office: +46 (0)10-615 6020 Mobile: +46 (0)70-815 1696 roger.oberholtzer@ramboll.se ________________________________________ Ramböll Sverige AB Krukmakargatan 21 P.O. Box 17009 SE-104 62 Stockholm, Sweden www.rambollrst.se -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Hello, On Aug 28 10:07 Per Jessen wrote (excerpt):

Actually, when a group of openSUSE users have needs that are not fulfilled by openSUSE, they might vote with their feet.

Of course! openSUSE is a free project and its users can do what they like (within the project's policies). I don't know if "voting with the feets" is excatly within the project's policies but I assume there exists some kind of voting mechanism in openSUSE to make decisions ;-) Kind Regards Johannes Meixner -- SUSE LINUX GmbH - GF: Felix Imendoerffer, Jane Smithard, Graham Norton - HRB 21284 (AG Nuernberg) -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Am 28.08.2015 um 17:40 schrieb Jim Henderson:

On Fri, 28 Aug 2015 10:23:58 +0200, Johannes Meixner wrote:

...

I don't know if "voting with the feets" is excatly within the project's policies but I assume there exists some kind of voting mechanism in openSUSE to make decisions ;-)

Maybe the phrase doesn't translate - "voting with your feet" means that when something doesn't meet your needs, you go somewhere else to get something that does.

What Per is saying is that if people who want x86 support from openSUSE don't get it, they'll find a distro that does.

Which, if the stats Richard has are accurate, means we might lose, what, a dozen users? ;) (Exaggeration for effect)

Jim

This poll is not representative but might give an answer to the question about user numbers: http://distrowatch.com/weekly.php?myaction=SeeVote&issue=20150629#poll btw. I'm still using 32bit hardware every day. Hendrik -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Not really, because it's not openSUSE-specific.

The actual download numbers for 32-bit openSUSE would be better guidance.

no , is same across whole world On 28 August 2015 at 18:21, Jim Henderson wrote:

On Fri, 28 Aug 2015 18:03:10 +0200, Hendrik Woltersdorf wrote:

...

This poll is not representative but might give an answer to the question about user numbers:

Not really, because it's not openSUSE-specific.

The actual download numbers for 32-bit openSUSE would be better guidance.

Jim

-- Jim Henderson Please keep on-topic replies on the list so everyone benefits

-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Fri, 28 Aug 2015 18:23:43 +0200, Ondřej Súkup wrote:

...

Not really, because it's not openSUSE-specific.

The actual download numbers for 32-bit openSUSE would be better guidance.

no , is same across whole world

If openSUSE users aren't downloading 32-bit in sufficient volume, that's the more useful measurement. Personally, I don't care how many people around the world use 32-bit systems. I care about the number of people who use openSUSE who use 32- bit systems, and that number, according to Richard, has been shrinking, even though 13.x provides a 32-bit version. The decline - and rate of decline - is the important measure. Supporting a 32-bit platform until nobody is using 32-bit platforms around the world is a demonstrably silly idea. That said, though, if there's a group of users who want to continue to build a 32-bit version, that's cool. Let it be like the Pi versions, or the other non-official builds like Evergreen or the education version. That's what open source is all about. I got rid of my last 32-bit system many years ago. When I was laid off, I had to scrape together enough money to buy a laptop, and I bought a 64- bit system for a pretty good price. I wouldn't say I was in poverty, but I certainly had to watch every penny I spent. My oldest system here is 10 years old; it's a 64-bit AMD system that maxes out at 2 GB of RAM - the system board won't support more. We go through this every time an architecture becomes obsolete - it happened with i386 support was dropped (an easy decision since the kernel dropped support for it), but there were a handful of people who said "but *I* still use an 80386 system! What about me?" - using a system that's out of date limits your options in a lot of ways. We also can't use old microchannel hardware any more, or EISA-bus cards. That's life in the technology space - things change, and you either keep up with the times, or you find your choices limited. I don't say that to be "mean" or anything - it's just the reality of the situation. If you have a platform that accounts for 90% of use and one that accounts for 20%, you don't allocate equal resources to the platform that has 10% - that's a poor way of managing resources. Jim -- Jim Henderson Please keep on-topic replies on the list so everyone benefits -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 08/28/2015 06:59 PM, Martin Pluskal wrote:

Simple bugzilla search shows that this year, following numbers of issues were reported for Tumbleweed (and had architecture specified) x86_64: 233 i386, i586, i686: 6

Note that although number of repors for TW is much larger (878), architecture is not specified in most cases, but ratio suggests that amount of 32-bit systems is bellow 3%.

Cheers

Martin Pluskal

And anyway, the more/less there are people "reporting 32 bits bugs" (to use Martin figures, thanks a lot Martin ! ), the more/less there will be people contributing to OpenSUSE-32. Dsant, from France -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Dne 28.08.2015 v 20:59 Greg Freemyer napsal(a):

On Fri, Aug 28, 2015 at 12:59 PM, Martin Pluskal wrote:

...

+1 Simple bugzilla search shows that this year, following numbers of issues were reported for Tumbleweed (and had architecture specified) x86_64: 233 i386, i586, i686: 6

Shocking! Bleeding Edge Tumbleweed users aren't using 32-bit. I'm amazed.

Seriously, give the 13.1 and 13.2 number over their lifetimes.

Well since you are so pleasant: for 13.1 it's 1432 vs 169 for 13.2 it's 808 vs 42 M

On Fri, Aug 28, 2015 at 3:16 PM, Martin Pluskal wrote:

Dne 28.08.2015 v 20:59 Greg Freemyer napsal(a):

...

On Fri, Aug 28, 2015 at 12:59 PM, Martin Pluskal wrote:

...

+1 Simple bugzilla search shows that this year, following numbers of issues were reported for Tumbleweed (and had architecture specified) x86_64: 233 i386, i586, i686: 6

Shocking! Bleeding Edge Tumbleweed users aren't using 32-bit. I'm amazed.

Seriously, give the 13.1 and 13.2 number over their lifetimes.

Well since you are so pleasant: for 13.1 it's 1432 vs 169 for 13.2 it's 808 vs 42

M

Martin, I apologize. I was a little flabbergasted you ran 32-bit numbers against tumbleweed since it seemed the least likely to have 32-bit users. So its about 12% for 13.1 and 5% for 13.2 Not sure what it really tells us, but there are people reporting bugs against 32-bit. Greg -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Carlos E. R. composed on 2015-08-29 02:09 (UTC+0200):

On 2015-08-28 22:44 (UTC-0400), Greg Freemyer wrote:

...

So its about 12% for 13.1 and 5% for 13.2

...

Not sure what it really tells us, but there are people reporting bugs against 32-bit.

Maybe it means that the 32 bit platform has less bugs than the 64 bit platform. Thus we should use 32 bit version, it is far better ;-) :-P

:-D Maybe it means 32 bit users aren't interested in latest and greatest things generally, and/or more interested in not needing to upgrade to a newer distro version when support terminates for their installed release. After all, 13.1 has scheduled to become Evergreen for quite some time. -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 On 2015-08-29 09:22, Martin Pluskal wrote:

Dne 29.08.2015 v 4:12 Felix Miata napsal(a):

...

Maybe it means 32 bit users aren't interested in latest and greatest things generally, and/or more interested in not needing to upgrade to a newer distro version when support terminates for their installed release. Then they are on their own - and I don't see what does amount of users using unsupported releases has to do with not building 32-bit leap and moving 32-bit Tumbleweed to ports.

You misunderstood. You removed an important sentence from his paragraph:

...

After all, 13.1 has scheduled to become Evergreen for quite some time.

Which is true, my 32 bit servers are on Evergreen. - -- Cheers / Saludos, Carlos E. R. (from 13.1 x86_64 "Bottle" (Minas Tirith)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) iF4EAREIAAYFAlXhqhAACgkQja8UbcUWM1yRfwD/dMcprwRWrsX55gXiH2FaYzB0 V/IlGYg8L8A4qPdMVg0A/0K6vfqD/SYCkKXZzJ+RiCZX/VT6J/t79iP2oMTYLOTz =bfmv -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 On 2015-08-31 17:46, Brian F. Yulga wrote:

I completely understand that it is not practical to continue official 32-bit support due to the general direction of the openSUSE project with Leap. Is there a way to make unofficial 32-bit support (e.g. a spin) less time or effort... for example, require a successful build against the "minimal x-window" install pattern, instead of all packages? (I have no idea because I haven't used OBS - I need to learn!)

I fear it will be impossible. We can not build it on susestudio because there will be nothing to base it on, no repos to draw from. And I don't know if it is possible to do an install disk on the OBS, if the 32 bit target is removed. We would have to do it all. Absolutely all. - -- Cheers / Saludos, Carlos E. R. (from 13.1 x86_64 "Bottle" (Minas Tirith)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) iF4EAREIAAYFAlXk41MACgkQja8UbcUWM1wxCQEAms5MwC4D2jxN+az5mbsivN28 cXEgoiX1Ne/3zRXK55gA/jxWCczs8riuu4+LSJdSU3B/BlV4lhqjhh0glT6DRlv7 =lkpj -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 On 2015-09-01 01:42, Malcolm wrote:

Hi Huh? You just add the OBS repo to add files/patterns to your SUSE Studio build target...

But there will not be any repo to add. They won't have any 32 bit version. - -- Cheers / Saludos, Carlos E. R. (from 13.1 x86_64 "Bottle" (Minas Tirith)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) iF4EAREIAAYFAlXk79gACgkQja8UbcUWM1zDcAD+MEyA+NaKUO6WotdKrgc/wRzg 3VEyvg7BMgqqmjA2C9sA/RLuRvXAOZhPbEkuiEsISvtb+I1tfjGt+QhJWWQxc3Wl =fOZC -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

"Carlos E. R." writes:

On 2015-09-01 01:42, Malcolm wrote:

...

Hi Huh? You just add the OBS repo to add files/patterns to your SUSE Studio build target...

But there will not be any repo to add. They won't have any 32 bit version.

The i586 architecture won't be removed from the repository, since it is still needed to build the -32bit packages. Andreas. -- Andreas Schwab, SUSE Labs, schwab@suse.de GPG Key fingerprint = 0196 BAD8 1CE9 1970 F4BE 1748 E4D4 88E3 0EEA B9D7 "And now for something completely different." -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 On 2015-09-01 02:56, greg.freemyer@ wrote:

Carlos,

OB$ builds various "ports" of opensuse: multiple ARM and PPC architectures. One of those is built via qemu on the Intel servers.

If 32-bit is demoted to ports status, I fully expect it to stay on OBS, just with a lower build priority than it has now. That is unlikely to be the problem. Further, OBS has integrated kiwi support, so building ISOs should still be possible.

I see.

The bigger issue is who are the enthusiasts that will maintain grub, grub2, lilo, initrd, and of course the kernel.

True.

There are arm enthusiasts that make sure it works on various arm platforms. Maybe they can inform us how much work it is? As I've said, i don't need an up to date 32-bit version of openSUSE, so I won't be one of those enthusiasts.

I might be, but I don't have the knowledge at all. I'm a developer by training, yes, but not in Linux. And certainly nothing about OBS, and some of what I thought I knew was/is wrong. - -- Cheers / Saludos, Carlos E. R. (from 13.1 x86_64 "Bottle" (Minas Tirith)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) iF4EAREIAAYFAlXlZKkACgkQja8UbcUWM1yeOAD9FFO8yIeWZmjZXGqObURbjbyl TYhCvs1Gs6ul49uU63EA/34c9udy8WMOHS/0SBViJTLfkSLn81de/X3AaoUmTnPS =9LAL -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 On 2015-09-01 11:04, Felix Miata wrote:

greg.freemyer@ composed on 2015-08-31 20:56 (UTC-0400):

...

The bigger issue is who are the enthusiasts that will maintain grub, grub2, lilo, initrd, and of course the kernel.

Seems to me decade or older 32 bit CPUs ought to get along just fine without any "maintenance" on Grub Legacy, and for most need none of what Grub2 provides that Grub Legacy doesn't.

Well, yes, unnecessary things would have to be dropped, and keep only the easiest to maintain. UEFI: no. There should not be UEFI/32 bit machines. lilo, grub 1: no. Grub2 is easier to maintain (not as user, but as maintainer). initrd? Keep dracut, I guess. Based in 13.2/tumbleweed. - -- Cheers / Saludos, Carlos E. R. (from 13.1 x86_64 "Bottle" (Minas Tirith)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) iF4EAREIAAYFAlXlcOkACgkQja8UbcUWM1yuyAD/X4WyAml0b5Jh1wAyBDnNRMO7 ZXrtWRVrIpQ6AiKMzuMA/ROmn/WJEW9q3IQJ11MyntyRehZUNv2I/vDSUL8+C7oo =eNTg -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 On 2015-09-01 12:58, Yamaban wrote:

On Tue, 1 Sep 2015 11:33, Carlos E. R. wrote:

...

UEFI: no. There should not be UEFI/32 bit machines.

Apple laughts at you. They started using EFI on 32bit CPUs. They have even the ugly construct of 32bit EFI on 64bit CPU boards.

Crumbs. :-} Consider that things have to be ported from the existing Leap. Anything not maintained there can't be used on 32 bits either, even less by people not experienced in making a distro... - -- Cheers / Saludos, Carlos E. R. (from 13.1 x86_64 "Bottle" (Minas Tirith)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) iF4EAREIAAYFAlXlljEACgkQja8UbcUWM1wGWAEAh4XZFOMMJ6naYLvthYn6nzb8 dSVMOANQw3KMRysHaEcA/Ri1x2aLefO+fGXj6HPTqycze/JjR9CKXQoRuYXJIkN+ =OUzt -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Tue, 01 Sep 2015 12:58:48 +0200, Yamaban wrote:

Did lilo (and syslinux) need any maintainance in the last 3 years?

I remember seeing something within the past day or so that lilo development is finished. Ah, yes, here: https://lilo.alioth.debian.org/ Jim -- Jim Henderson Please keep on-topic replies on the list so everyone benefits -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Roman Bysh composed on 2015-09-01 19:23 (UTC-0400):

Yes. 2015 is the last year for lilo. Ah..Those were the pre grub1 days of lilo.Very easy to work with.

Only as long as lilo.conf never had a typo, and lilo always got remembered to be run after an edit to lilo.conf. Grub's on the fly cmdline editing was a welcome difference. I've never looked back. -- "The wise are known for their understanding, and pleasant words are persuasive." Proverbs 16:21 (New Living Translation) Team OS/2 ** Reg. Linux User #211409 ** a11y rocks! Felix Miata *** http://fm.no-ip.com/ -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Отправлено с iPhone

2 сент. 2015 г., в 4:20, Roman Bysh написал(а):

...

On 09/01/2015 07:57 PM, Felix Miata wrote: Roman Bysh composed on 2015-09-01 19:23 (UTC-0400):

...

Yes. 2015 is the last year for lilo. Ah..Those were the pre grub1 days of lilo.Very easy to work with.

Only as long as lilo.conf never had a typo, and lilo always got remembered to be run after an edit to lilo.conf. Grub's on the fly cmdline editing was a welcome difference. I've never looked back. Exactly. You also had to run lilo -v for updating. I always liked Grub1 as it was easy to understand. Whereas Grub2 relies on scripts.

No, it does not. It offers "easy configuration" tool grub-mkconfig for those who do not want to write own grub.cfg. But this is entirely optional. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Monday 31 of August 2015 10:46:30 Brian F. Yulga wrote:

Agreed, writing as a long-time user that does not need latest and greatest all the time! My ASUS Eee PC Tablet -- purchased new in 2011 -- is 32-bit only. But, it has (after upgrades) an SSD and 2 GB of RAM, and ran most openSUSE releases since 11.4. A four year-old netbook that I use regularly doesn't feel that old or obsolete to me, especially when running Linux.

What model are you talking about? I tried Google and found "ASUS EEE Pad Transformer TF101" which has NVidia Tegra 3 CPU, i.e. ARM. That wouldn't be relevant for the discussion about i586 architecture. Michal Kubeček -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Tue 01 Sep 2015 08:27:38 AM CDT, Brian F. Yulga wrote:

On 2015.08.31 23:56, Michal Kubecek wrote:

...

On Monday 31 of August 2015 10:46:30 Brian F. Yulga wrote:

...

Agreed, writing as a long-time user that does not need latest and greatest all the time! My ASUS Eee PC Tablet -- purchased new in 2011 -- is 32-bit only. But, it has (after upgrades) an SSD and 2 GB of RAM, and ran most openSUSE releases since 11.4. A four year-old netbook that I use regularly doesn't feel that old or obsolete to me, especially when running Linux. What model are you talking about? I tried Google and found "ASUS EEE Pad Transformer TF101" which has NVidia Tegra 3 CPU, i.e. ARM. That wouldn't be relevant for the discussion about i586 architecture.

Michal Kubeček

This one has an Intel Atom CPU, and came with Win 7 Starter 32-bit (which I never used :-) Model number under the battery: T101MT-EU27-BK

http://www.asus.com/Notebooks/Eee_PC_T101MT/specifications/

-- Brian Y.

Hi And both of those cpu's are 64bit.... http://ark.intel.com/products/42503/Intel-Atom-Processor-N450-512K-Cache-1_6... http://ark.intel.com/products/55637/Intel-Atom-Processor-N570-1M-Cache-1_66-... You also realize that a valid windows product code works with 32 and 64bit, so you could have installed the 64bit starter if so inclined. -- Cheers Malcolm °¿° LFCS, SUSE Knowledge Partner (Linux Counter #276890) SUSE Linux Enterprise Desktop 12 GNOME 3.10.1 Kernel 3.12.44-52.10-default up 23:09, 4 users, load average: 0.71, 0.60, 0.49 CPU Intel® Core i3-3227U CPU @ 1.90GHz | GPU Intel® HD Graphics 4000 -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 09/01/2015 11:00 AM, Brian F. Yulga wrote: ????? -- Ken Schneider -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On 2015.09.01 09:02, Malcolm wrote:

On Tue 01 Sep 2015 08:27:38 AM CDT, Brian F. Yulga wrote:

...

On 2015.08.31 23:56, Michal Kubecek wrote:

...

On Monday 31 of August 2015 10:46:30 Brian F. Yulga wrote:

...

Agreed, writing as a long-time user that does not need latest and greatest all the time! My ASUS Eee PC Tablet -- purchased new in 2011 -- is 32-bit only. But, it has (after upgrades) an SSD and 2 GB of RAM, and ran most openSUSE releases since 11.4. A four year-old netbook that I use regularly doesn't feel that old or obsolete to me, especially when running Linux. What model are you talking about? I tried Google and found "ASUS EEE Pad Transformer TF101" which has NVidia Tegra 3 CPU, i.e. ARM. That wouldn't be relevant for the discussion about i586 architecture.

Michal Kubeček

This one has an Intel Atom CPU, and came with Win 7 Starter 32-bit (which I never used :-) Model number under the battery: T101MT-EU27-BK

http://www.asus.com/Notebooks/Eee_PC_T101MT/specifications/

-- Brian Y.

Hi And both of those cpu's are 64bit.... http://ark.intel.com/products/42503/Intel-Atom-Processor-N450-512K-Cache-1_6... http://ark.intel.com/products/55637/Intel-Atom-Processor-N570-1M-Cache-1_66-...

You also realize that a valid windows product code works with 32 and 64bit, so you could have installed the 64bit starter if so inclined.

Wow, you are absolutely right! The 64-bit capability was not well-advertised, and I did not know it was possible to use a 64-bit Win 7 when it came with 32-bit. My apologies. My other Eee PC from ~2009 is definitely 32-bit... https://www.asus.com/Notebooks/Eee_PC_S101/specifications/ http://ark.intel.com/products/36331/Intel-Atom-Processor-N270-512K-Cache-1_6... ...and using IceWM for primarily xterm and ssh. It's the first time I tried the "Minimal X Window" install with openSUSE, and have been pleased with the results. -- Brian Y. -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

Am 28.08.2015 um 20:59 schrieb Greg Freemyer:

On Fri, Aug 28, 2015 at 12:59 PM, Martin Pluskal wrote:

...

+1 Simple bugzilla search shows that this year, following numbers of issues were reported for Tumbleweed (and had architecture specified) x86_64: 233 i386, i586, i686: 6

Shocking! Bleeding Edge Tumbleweed users aren't using 32-bit. I'm amazed.

Seriously, give the 13.1 and 13.2 number over their lifetimes.

(I'd do it, but I don't know how to run those searches in bugzilla.)

I'm also curious what the breakdown is at SUSESTUDIO. I only publish one appliance there, but it is 32-bit to ensure the widest hardware compatibility, and yes that includes 10+ year old machines.

Greg -- Greg Freemyer www.IntelligentAvatar.net

Damn, I could have raised the i*86 number by 16%, if I had not forgotten to set the hardware in my last bug report :) My personal interest is running old hardware. I could afford to by new hardware, but I _hate_ to throw away a good working laptop with an excellent display just because it is "old" by someone's standards. So I raised my hand, to tell that I'd love to see 32-bit versions of openSUSE in the future. If it makes a difference, fine. If not, the world will keep on turning. Hendrik -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Friday, August 28, 2015 10:27:34 PM Per Jessen wrote:

Hendrik Woltersdorf wrote:

...

My personal interest is running old hardware. I could afford to by new hardware, but I _hate_ to throw away a good working laptop with an excellent display just because it is "old" by someone's standards.

So I raised my hand, to tell that I'd love to see 32-bit versions of openSUSE in the future. If it makes a difference, fine. If not, the world will keep on turning.

Personally speaking, I think you're the kind of user we want to keep hanging on. The amateurs who love what they're doing. (look up what amateur means if you don't know).

I also love to fire up an ancient box just to see if the latest openSUSE will install and/or work.

Well, I will hate to throw away a bunch of "old" and excellent pieces of hardware just because useful OS Standards changed. I will cry loud with no remedies to keep those modular and good machines "before spy era" ;-D At the end, will say "Farewell my beloved ones, we had a good time together. Now it's time to switch for the sake of new hardware generation" :-) Regards, Rick -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 On 2015-08-29 05:17, Felix Miata wrote:

I specified either PC, i586, 32 bit, or x86 on 6 of the 23 BOO bugs I filed in the past 365 days. Those 6 are either 32 bit exclusive or 32 bit inclusive. Likely half the rest, not counting opensuse.org infrastructure bugs, would have been discovered here on a 32 bit boot, possibly confirmed on a 64 bit boot. 13 of the 23 are currently on product Factory. More would have been filed had I not made the effort to determine what I found was upstream sourced and filed there instead.

Well, it is true. When one has a mixture of 32 and 64 bit machines, and one finds a bug that is present in several, then one doesn't need to specify that the bug is 32 bit specific. The amount of bugzillas entered with "32bit" marked in the field is thus not a real indicator of how many 32 bit machines our userbase has. - -- Cheers / Saludos, Carlos E. R. (from 13.1 x86_64 "Bottle" (Minas Tirith)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) iF0EAREIAAYFAlXhrCoACgkQja8UbcUWM1wiggEAm/KeLc/OBtLZQvojchclV7Kn wkk3zaBCi2lopw9Q5y0A9RTSZAQnloOYckwhQIcuT6JvT8s7iwLGYJzgeOIHs1Q= =7IEF -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Fri, 28 Aug 2015 20:38:12 +0200, Per Jessen wrote:

...

That said, though, if there's a group of users who want to continue to build a 32-bit version, that's cool. Let it be like the Pi versions, or the other non-official builds like Evergreen or the education version. That's what open source is all about.

Where do I/we start? It seems to be slightly different to both Evergreen and the education version as they are based on existing versions. (to my knowledge).

If it were me, I'd start by learning how to build the packages and build a release. That seems like a logical first step.

...

I got rid of my last 32-bit system many years ago. When I was laid off, I had to scrape together enough money to buy a laptop, and I bought a 64-bit system for a pretty good price. I wouldn't say I was in poverty, but I certainly had to watch every penny I spent.

Apologies, I may be stepping on somebodys toes, but it's really not so much about running on real 32-bit hardware. The 32-bit-only hardware _is_ dying, but it doesn't mean the software has to die too. For the purpose of this discussion, perhaps we ought to focus on running 32-bit systems on 64-bit hardware. (that's my current focus any way).

To what end? Why would you install a 32-bit OS on 64-bit hardware? If it's to deal with some deficiency in the 64-bit version, fix the deficiency rather than maintain an entire infrastructure to build a 32- bit release. Jim -- Jim Henderson Please keep on-topic replies on the list so everyone benefits -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Fri, 28 Aug 2015 22:23:07 +0200, Per Jessen wrote:

Jim Henderson wrote:

...

On Fri, 28 Aug 2015 20:38:12 +0200, Per Jessen wrote:

...

...

That said, though, if there's a group of users who want to continue to build a 32-bit version, that's cool. Let it be like the Pi versions, or the other non-official builds like Evergreen or the education version. That's what open source is all about.

Where do I/we start? It seems to be slightly different to both Evergreen and the education version as they are based on existing versions. (to my knowledge).

If it were me, I'd start by learning how to build the packages and build a release. That seems like a logical first step.

I agree, sounds good. I tried joining the release engineering team back in June, but my attempts were so far ignored. Is there documentation available anywhere? I know how to build packages, now how about an entire release?

Richard? Can you hook Per up?

...

...

...

I got rid of my last 32-bit system many years ago. When I was laid off, I had to scrape together enough money to buy a laptop, and I bought a 64-bit system for a pretty good price. I wouldn't say I was in poverty, but I certainly had to watch every penny I spent.

Apologies, I may be stepping on somebodys toes, but it's really not so much about running on real 32-bit hardware. The 32-bit-only hardware _is_ dying, but it doesn't mean the software has to die too. For the purpose of this discussion, perhaps we ought to focus on running 32-bit systems on 64-bit hardware. (that's my current focus any way).

To what end? Why would you install a 32-bit OS on 64-bit hardware?

Jim, people (and myself) have pointed that out earlier in this thread:

running 32bit apps running 32bit virtual guests.

Both use less memory in 32bit.

Ah, "scarcity". Not a really good argument IMHO. I've got one system here with 8 GB of RAM, one with 16 GB of RAM, and one with 32 GB of RAM. I run 64-bit SLE in VMs on the latter, typically about 8 at a time.

...

If it's to deal with some deficiency in the 64-bit version, fix the deficiency rather than maintain an entire infrastructure to build a 32- bit release.

The deficiency is that a 64bit app uses more memory than a 32bit ditto.

How much? 2x? 1.5x? 16 bytes more because of double-byte pointers? It can't be *that* significant, certainly not significant enough to justify the resources to build *an entire 32-bit distribution* to deal with a handful of people who still treat memory is a scarce resource. Jim -- Jim Henderson Please keep on-topic replies on the list so everyone benefits -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

On Sat, 2015-08-29 at 02:26 +0200, Carlos E. R. wrote:

On 2015-08-29 00:39, Jim Henderson wrote:

...

On Fri, 28 Aug 2015 22:23:07 +0200, Per Jessen wrote:

...

...

The deficiency is that a 64bit app uses more memory than a 32bit ditto.

How much? 2x? 1.5x? 16 bytes more because of double-byte pointers?

I remember a similar discussion long ago. We did some testing with sample programs, and some would use double memory when compiled for 64 bit than when compiled for 32 bit, and not using pointers at all.

Of course, it is a problem with the program design. You can declare some type of integer array, which /may/ compile with different integer sizes (16, 32, 64 bits...) depending on the architecture. Simply because the default for the used declaration is "biggest word available" (which was not a problem when "biggest" meant "32 bit").

Of course, you can declare a different type of integer and this is not a problem anymore.

Ok, but this was with "sample test programs", it can be said. True. But that thread originated because the OP had found a particular application (sorry, I don't remember which) which used huge amounts of memory, and in 64 bits it doubled the already huge amount.

Mind, I'm not saying that "any 64 bit install uses double the memory than a 32 bit install". No. I say that "it can" use up to that. What is the figure currently, I don't know. I haven't tested. I was told, I think, but I forgot.

-- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

s/32/16/g s/386/286/g Woah. It's 1987 all over again. :P -- James Mason Technical Architect, Public Cloud openSUSE Member SUSE jmason@suse.com ------------------------------------- SUSECon 2015: Register at susecon.com

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 On 2015-08-29 08:17, Cristian Rodríguez wrote:

On Fri, Aug 28, 2015 at 9:26 PM, Carlos E. R. wrote:

...

Ok, but this was with "sample test programs", it can be said. True. But that thread originated because the OP had found a particular application (sorry, I don't remember which) which used huge amounts of memory, and in 64 bits it doubled the already huge amount.

Sample programs. or even single real-life programs are meaningless in this context (unless there is a memory leak or a design issue).. what matters is whole-system performance/memory usage and the tradeoffs involved..

Well, when I have time, I can do identical 32 and 64 bit installs, under vmware, and compare memory usage with identical loads. What I have said here is that I know that there are applications that use double. - -- Cheers / Saludos, Carlos E. R. (from 13.1 x86_64 "Bottle" (Minas Tirith)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) iF4EAREIAAYFAlXhrNYACgkQja8UbcUWM1zNVwEAkk1IfRNEvegGkD/rsJxfVkXk OIIy9DWasd86ZiOL7BsA/RxN3sD95CEF2Fn54dQKTcO/HfsMENdl5wZki+UUzO7k =EYH6 -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA256 On 2015-08-29 20:03, Cristian Rodríguez wrote:

On Sat, Aug 29, 2015 at 10:00 AM, Carlos E. R. wrote:

...

Well, when I have time, I can do identical 32 and 64 bit installs, under vmware,

Don't..at least not under vmware..thing we do not know what is doing and can't fix.. use KVM instead.

Sorry, no. I don't know nor use KVM :-) - -- Cheers / Saludos, Carlos E. R. (from 13.1 x86_64 "Bottle" (Minas Tirith)) -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.22 (GNU/Linux) iF4EAREIAAYFAlXh/sgACgkQja8UbcUWM1wXhwD/b6mCYk+YYd0jqBO+7CW9MpqC YX9Lbu75rjBnoxDitX4A/3CoEa+Z28C0tDlDhh31D2fubT//UOjrGfEJmlwPG0K6 =H9mW -----END PGP SIGNATURE----- -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org