Page 2 of 2

Posted: Wed 25 Dec 2019, 14:28
by rcrsn51
So when the mainstream distros decide to abandon 32bit, they don't know what they are doing?

Posted: Wed 25 Dec 2019, 15:38
by s243a
rcrsn51 wrote:So when the mainstream distros decide to abandon 32bit, they don't know what they are doing?
I think that developers often run better hardware than users and this might skew their views.

Posted: Wed 25 Dec 2019, 17:35
by Terry H
The thing with creating newer 32 bit releases and discussing using them on old hardware, is that people aren't using these newer releases. Seems that everyone is using the older releases and don't even consider using new distribution releases. So as long as forum members are able to keep the older releases functional, people will be able to keep older equipment working.

Posted: Wed 25 Dec 2019, 17:56
by musher0
I say we Puppyists keep publishing 32-bit versions to prevent sending good boxes to the
land fill.

Not mentioned yet, an additional argument s243a brought forward a while back: 32-bit
CPUs do not have the "predictive" whatchamacallit that opens up 64-bit CPUs to malware.

Posted: Wed 25 Dec 2019, 19:27
by bigpup
Get back to the question.

What other Linux OS repositories should be used for the core Linux files and programs???

What Linux kernel should be used????

Should it be PAE or not PAE????

What about specific Puppy software that should be in it???
This is important to me.
There is a lot of very good Puppy specific programs, out there, that really need to be part of a Puppy version.
Example:
cpucurtemp (CPU temp tray display)
http://murga-linux.com/puppy/viewtopic.php?t=115684
FrugalPup installer (do frugal installs with options for boot loader)
http://www.murga-linux.com/puppy/viewtopic.php?t=114340

Posted: Wed 25 Dec 2019, 19:41
by s243a
bigpup wrote:Should it be PAE or not PAE????
Ideally both but it is really up to whatever the puppy builders want to build.

Posted: Wed 25 Dec 2019, 20:04
by bigpup
s243a wrote:
bigpup wrote:Should it be PAE or not PAE????
Ideally both but it is really up to whatever the puppy builders want to build.
True, but it does affect what processors will be supported.
The processor has to have PAE support feature to use a PAE Linux kernel.
Back to Pentium M or older processors not having support for PAE.

How much memory will be available to use.
No PAE - it will see up to about 2GB.
Have more than 2 GB RAM.
Example:
4GB of RAM
you will need to enable PAE to see all 4GB of RAM. Whilst a 32Bit CPU should in theory be able to use up to 4GB of RAM without PAE, the problem with this is that it requires all 32bits of the address bus to use the 4GB, leaving nothing left for things like graphics cards.

Posted: Wed 25 Dec 2019, 20:15
by musher0
bigpup wrote:Get back to the question.
(...)
I did answer: "Should there be a 32-bit UPupFF in 04/2020?"
Bigpup, you are splitting hairs with your sub-questions.
The specific components and choices should be left to the developers.

Posted: Thu 26 Dec 2019, 01:30
by Mike Walsh
musher0 wrote:
bigpup wrote:Get back to the question.
(...)
I did answer: "Should there be a 32-bit UPupFF in 04/2020?"
Bigpup, you are splitting hairs with your sub-questions.
The specific components and choices should be left to the developers.
I would dispute the 'hair-splitting' allusion. There is actually a sound point being made here, musher0.

If there are eventually no 32-bit OSs being produced any longer (by any Linux distro), then we'll be back to building Puppies the way BarryK had to originally; one item at a time, and compiling absolutely everything as you go, from the ground up.

As do the FatDog team, I believe, using Linux from Scratch.....? It may not be immediately, but it is coming, sooner or later.

So the choice will be something like

a ) Continue to build Pups the way we do now (using available 64-bit components.....i.e., relatively easily). Or
b ) Attempting to build 32-bit systems the hard way.....having to compile everything.

I know some of us like compiling.....but there are limits, mate!


Mike. :wink:

Posted: Thu 26 Dec 2019, 03:18
by musher0
Hello all.

Out of curiosity, I did a search for 32-bit Linuxes on distrowatch and found that there
are one hundred and thirty two ( that's right: 132 ) 32-bit distros active.

This is how I searched:
I went to https://distrowatch.com/search.php#advanced.
In the more elaborate search section (not the top one), I ticked "Linux" and further
below all the "i?86" boxes, and then I clicked on the query button.

I don't drink alcohol, but maybe my eyes are playing tricks on me?
Please redo this exercise, and confirm or contradict! :)

~~~~~~~~~~~~~~
Question @peebee:
would switching to the trisquel 32-bit repo be of any help for a future 32-bit upup
derivative?

The Trisquel distro is mentioned in the distrowatch list above and also in this (strangely
enough) ubuntu compilation of 32-bit distros.

~~~~~~~~~~~~~~

In short, 32-bit distros are not dead yet.

Posted: Thu 26 Dec 2019, 03:29
by musher0
As a note, reading
https://en.wikipedia.org/wiki/Physical_ ... _Extension
it seems that all CPUs built from 2003 onward can support PAE.

May I suggest that all new 32-bit Puppies be PAE-enabled?

In any case, the OS Puppy traditionally uses less than 1 Gb of
RAM, even with major apps running.

What would the rest of the RAM be used for? A RAMdisk?

Just a thought.

Posted: Thu 26 Dec 2019, 04:06
by darry19662018
May be a Linux from scratch Pup anyway like you say there are other 32bit distros like Void, Slitaz, Arch 32bit, Devuan, NuTyx..........................

Posted: Thu 26 Dec 2019, 12:01
by ozsouth
Peebee, that list of 'problematic' items is a huge challenge & the situation is not likely to improve.
You do so much already that I think you should take a less stressful option - maybe debian, as it will continue to be developed.
Alternatively, I note that Bionic (standard) EOL is April 2023. Puppy users with old machines/32bit preference could continue with that (it's very good). Perhaps just keep updating upupbb & review the situation in 3 years time? Of course there's always ScPup (32bit) that you're doing too.

Posted: Thu 26 Dec 2019, 14:05
by ndujoe1
I agree when you build a Puppy what is it is in it and operational it works.

I agree over time the PPM will not be useful and software updates will not work on the 32 bit system. As others have noted the 32 bit architecture will not support some of the new features of software.

I suppose you one would have to compile from source for any 32 bit software down the road.

But like many XP users if it aint broke don't fix it.

I switched to a 64 bit puppy do to the LInux date limitation coming down the pike.

Posted: Thu 26 Dec 2019, 19:02
by peebee
Many thanks everybody for the various comments - all greatly appreciated.

The conclusions I have reached are:

1. there are already "enough" 32-bit pups
2. an Ubuntu Focal Fossa based 32-bit pup (UPupFF) is not viable given the reduction in the Ubuntu 32-bit repo contents
3. an Ubuntu Focal Fossa based 64-bit pup is a possible candidate but this is down to @666philb, the 64-bit Ubuntu developer
4. both 32-bit and 64-bit pups based on Debian Bullseye might be feasible but this is down to @Radky and @josejp2424, the Debian Stretch & Buster developers
5. there is a possible "gap-in-the-market" for a 32-bit Trisqel Etiona 9.0 based T9Pup but this is not a high priority and other than being "different" does not meet any real need

I will therefore try to keep BionicPup32 (UPupBB) and UPupEE (and LxPupBionic) up-to-date until Ubuntu removes support at EOL.
ScPup32 and ScPup64 will be my JWM/Rox pups and will transition to Slackware-15 if and when this is released.
LxPupSc32 and LxPupSc64 (LXDE pups) similarly.

Cheers
PeeBee

Posted: Thu 26 Dec 2019, 19:17
by musher0
"Tears", rather than "Cheers", I'd say...

But it's your decision, and thanks for the wonderful work you've been doing
during all those years.

BFN,