excalibur dist-upgrade?

Altoid · 2025-08-15 13:37:11

Hello:

As I have mentioned a few times, these days my main box runs on Devuan 5 / Daedalus, which I update and back-up once a week to a recovered WD-MBL running on OpenWRT:

$ uname -a
Linux devuan 6.1.0-38-amd64 #1 SMP PREEMPT_DYNAMIC Debian 6.1.147-1 (2025-08-02) x86_64 GNU/Linux
$

No issues save those which may have appeared while doing some housekeeping, more than anything to clean up stray matter / files / unnecessary packages left over from my system's dist-upgrade path beginning with ascii or Jesse. As it has been a while, I cannot exactly recall when, but it must have been ~ 05/2017.

I have recently seen quite a few posts from members (some quite eager) to do an upgrade to Devuan 6 / Excalibur, which I expect must be for a variety of valid reasons.

As for me, I cannot at least for the time being, find one to move from Daedalus, much the same way I was not able to find one to move from Beowulf with which I had no issues and was quite comfortable with (but was being mothballed) to Chimaera, process which was nothing short of a train-wreck for which I hold XFCE mainly responsible.

A last resort effort by way of yet another dist-upgrade to Devuan 5 / Daedalus straightened most everything out and thus avoided a tiresome roll-back or clean installation.

My box is a fully upgraded ca. 2008 BIOS boot only (ie: no secure boot crap) Sun Ultra24 workstation with an early version of the infamous Intel MEI which I keep at bay (?) by blacklisting the respective modules, and an Intel Q9550 CPU with 8.0Gb RAM which has been running perfectly well on Linux since late 2015 and on Devuan Linux since ~ 05/2017.

The MB has PCIe v2 / PCI v2.3 slots, USB2.0 controller and GbE.
Twin Quadro FX 580 cards feeding three 19" monitors run (well enough) with the latest nouveau driver and the LSI SAS1068E and the 2940U/UW SAS/SCSI controllers have worked perfectly well from the very first Linux installation.

So it would seem that (?) there is nothing that would require much of a dist-upgrade as I expect that any and all applicable Linux drivers / modules* are quite mature by now.

That said, I have no plan (or willingness) to spend my pension money on new (unneeded) hardware as everything works as expected.
My only (small) peeve is that the nouveau driver could do much better at running a 3x monitor desktop but that is about it.

In the 10+ years I have been running on Linux I have seen the size of each release grow and grow and cannot but wonder what more is in store.
eg:
Devuan Chimaera netinstall *.iso: 372.00 MB - UEFI installer: 0.754 MB
Devuan Daedalus netinstall *.iso: 477.80 MB - UEFI installer: 23.00 MB
30X more code was added to the UEFI partition on the road between Chimaera and Daedalus.

Frankly, I do not think I can expect anything in the way of improvements for my hardware.

And after reading this I cannot but to agree with this note:

## Old Business
- Debian Trixie has been released
- And what a cluster-f*** it is!

Not that I am wanting to do a dist-upgrade as I will probably be able to get by (like I did with Beowulf) by using a backported kernel if necessary but eventually the time will come to move forward albeit with the bare metal in my box in perfect working condition and unchanged, save maybe larger/faster drives or newer monitors.

I guess that, besides @greenjeans and I, there may be more members in a similar situation from which I would like to hear an opinion.

Best,

A.

* corrections welcome.

Last edited by Altoid (2025-08-15 15:50:23)

RedGreen925 · 2025-08-15 15:04:18

Not that I am wanting to do a dist-upgrade

Doubt you would want to the recommended command is now apt full-upgrade after using the normal apt update then apt upgrade to ensure you are using the latest software before the move to the new OS. For the supposed mess upon upgrading. I did a Excalibur upgrade on one of my servers on cloned drive to test it out couple of days ago. It went perfectly fine so I cloned back the drive to it and two other server machines that were running the Daedalus. I have never really had problems with the old dist-upgrade or the new full-upgrade in twenty years of using a Debian based system. Whenever Woody came out is when I started so the next Debian 4 would have been my first experience doing it. They usually just work for me and the vast majority of people out there. Some have problems but that is to be expected as some have made so many changes to the defaults there are bound to be conflicts.

golinux · 2025-08-15 15:26:32

I am going to post another link here because it reminds me of the Trixie release notes. I have skimmed that long and detailed oeuvre but not gotten too deep. It seems that Wayland is just more fruit from the poisoned tree:

https://stoppromotingwayland.netlify.app/

Perhaps it is time to put in more effort to ensure that xorg remains a viable option. Time to step up, folks!

g4sra · 2025-08-15 16:55:10

After all the bad news I have read about trixie I will be postponing dist-upgrade (or full-ugrade).

I have no need or interest in a distro optimised for cloud (AWS etc) or hosting dozens of VM's.
I was hoping for a change in leadership and a board moving away from commercialisation of the distro, but it looks like IBM\Redhat have won.
At some point I will need to break away from Debian as the upstream for technical reasons, and unfortunately that will mean Devuan too.
For the long term there is only one possible way out I can see, the freezing of the distro (here before 'usrmerge') as 'Trinity Desktop Environment' did with KDE3.
History is a great teacher, if you think 'usrmerge' is a good idea then go research the reason for the split in the first place (Bell Labs, AT&T).

I do not understand this craze for constantly replacing reliable efficient software for newer software which brings nothing to the table other than being newer and doing something differently (not better).
New kernels (security fixes and new modules for new hardware) will be required from time to time, and the occasional new userspace app to manage them, but that is it.
To me that looks like the perfect distro model for "Continuous Integration" or if you prefer the term, "Rolling Release"

"If it ain't broken don't fix it!"

Only use resources for fixing what is broken, not willy-nilly changing of packages.

Camtaf · 2025-08-15 17:05:17

@Altoid

I'm an oldie too, & I upgrade my computers when MS Windows users sell their old kit because it can no longer support MS's latest offerings....

One of my older laptops is a 2011 model i3, 2.13GHz with 4GB ram, & I have just upgraded the disk to an M2 SATA SSD in an adapter, still more than adequete, & running a slightly modified MX Linux install, (which is 'heavier' than Devuan/XFCE).

Last edited by Camtaf (2025-08-15 17:19:19)

greenjeans · 2025-08-15 17:15:29

@g4sra, well said!! Rolling releases just seems like the thing for gluttons for punishment.

I have had good luck upgrading various Devuan and Vuu-do isos to excalibur up until a couple days ago when I had my first failure, barely even got started and it errored out.

To be honest i'm surprised it ever works, lol, that's a LOT of packages at once!

But yeah, I think i'm with brocashelm and g4sra on this one, looks like i'll be running daedalus for years to come, it's wonderful and stable and in reality should be even more so now that updates should be greatly lessened.

Camtaf · 2025-08-15 17:17:35

@g4sra

Maybe take a look at Void Linux as a possible alternative - https://voidlinux.org/

RedGreen925 · 2025-08-15 23:09:37

Maybe take a look at Void Linux as a possible alternative - https://voidlinux.org/

I tried it once and you had better be prepared to be doing a lot of searching on the way to do a lot of things Debian does for you by default. Starting with booting it on another machine if cloning as their broken EFI implementation uses a direct entry to the firmware and does not setup the normal /EFI/BOOT/BOOTX64.efi that allows you to use the boot menu to choose another install. To be fair the Arch people do the same. So you need to set that up yourself then if logging in via ssh good luck on knowing what machine you log into as it does no tell you with either the MOTD when logging in or give a clue in the Terminal title you are using if graphical. Then numerous other things I had to go and copy the sensible Debian default configs to it I forget now but are still laying around in that install notes files. Oh and now I think the total deal breaker for me was in KDE it would show existing bookmarks in Dolphin with absolutely no way to add, modify or delete them without going to the config file and doing it manually. Additional deal breaker totally non intuitive naming of the packages if you need to install and you will because the dependency resolution is trash, if use to the sensible Debian naming scheme. That all from the memory of that experience doing it that comes to mind quickly.

Camtaf · 2025-08-16 09:25:31

Void/XFCE worked out of the box for me, only had to add Gnumeric & PySolFC, if I remember.

(Total install about 3GB on disk)

I'm just a (desktop) user, internet, videos, music, image viewing, spreadsheet work.

Last edited by Camtaf (2025-08-16 09:26:52)

mweishaar · 2025-08-17 21:21:27

@Altoid, I am pretty much in the same boat as you, I've just been using Linux on my desktop since '98. Got to Devuan because Mint forced their init system on me, and the new one was giving me problems. I started on ASCII XFCE in 2018 and have dist-upgraded since, and also have a 3-monitor setup. I will hold off on upgrading Excaliber once I think it's feasible to do so, just as I have in past upgrades. All of mine have had minor issues. I do a nightly rsync of my home drive to a separate disk. I have switched over to an AMD GPU because they just work better w/Linux. If you can zoom in on my image you can see more of my system info if you're curious.
3-monitor-setup Devuan

Last edited by mweishaar (2025-08-17 21:21:58)

stargate-sg1-cheyenne-mtn · 2025-08-18 11:45:26

hat-tip to @Altoid for the somewhat hidden link to trixie issues:

https://www.debian.org/releases/trixie/release-notes/issues.html

hat-tip to @golinux for the wonderful link to wayland issues:

https://stoppromotingwayland.netlify.app/

kudos to @g4sra @Camtaf @greenjeans @RedGreen925 @mweishaar for additional insights!

re: Void Linux, please see:

https://unixdigest.com/articles/some-of-the-great-gnu-linux-distributions.html#void
specifically of note:

From a technical perspective, Void Linux is another great Linux distribution, but if you seek out Void Linux because you want a Linux distribution without systemd, you need to know that the Void Linux project is not hostile towards systemd. They actually used to have systemd, but removed it only because systemd doesn't support Musl. However, from what I understand, there are no future plans of going back to systemd.

also this as well:

https://unixdigest.com/articles/void-linux-a-great-and-unique-linux-distribution.html

still we must always consider the writing on the wall:

https://dev1galaxy.org/viewtopic.php?id=3074

UnixMan1230 · 2025-08-18 15:10:25

@g4sra and @golinux

Let us really, truly for a moment look at the overall environmental and economic cost of GNOME 3, shall we? This will probably not shock either of you, but for those who have never thought about this, it might be an eye opener.

* Unity
* Budgie
* Cinnamon
* MATE
* Pantheon
* CAFE
* LXQt (albeit indirectly)
* Linux Mint's XAPP Project

Unity: Ubuntu's original response to GNOME 3. Hundreds if not thousands of man hours poured into a brand new desktop environment, over at least 4 years. It stands to reason that if they've put at least 4k hours into this, at a minimum dev compensation of $25/hr, that's at least $100k right there. Possibly much more. And that does nothing to quantify the value of the volunteer hours poured into Unity as well, likely causing this figure to balloon even more. Then, of course, they dropped it when it wasn't profitable. Now a whole new community has spawned around it just to keep it alive, likely putting hundreds more hours into keeping it usable.

Budgie: Built directly on GNOME, but still divergent enough that they have consistently run into problems with theme and representation lock-in. Now, they are looking at having to rewrite Budgie atop KDE Frameworks and Qt in order to just have proper theming and a stable UI. This represents hundreds if not thousands of man hours as well, once all is said and done. It took over 24 hours for Joshua Strobl to rewrite one of his own apps in Kirigami/Qt, based on the timestamps of his streams. So with this figure in-hand, assuming he leads the charge a similar rate, this can climb up to 300-400 hours very easily. Assimung the same $25 rate from earlier, easily another $10k in the hole. At a *minimum*. This does not even count the cost of R&D to bring feature parity.

Cinnamon: Going on 12 years now, built and maintained by Linux Mint after they realized GNOME extensions alone were not going to cut it. Twelve years of continuous development, predominantly by Clement himself but far from exclusively so. Assuming even a quarter of Mint's monthly donations goes into Cinnamon, which are approximately $7-12K a month, that puts the total investment in Cinnamon, that puts the total investment at $288k, at a *minimum*.

MATE: The most direct response to GNOME 3 in this list, MATE is a fork of the GNOME 2 codebase. Still going to this day, much like Cinnamon, its direct costs likely include hosting, web domain registration, and of course the man-hours needed to maintain it. Even though this was probably the least work of everything on this list, its still not a small endeavor. At least 300 man hours had to have gone into this project in the beginning, and by now a conservative estimate would still put it at around 2K man hours at this point. Thats $50k on labor alone. Add in the webhosting costs, and that figure grows to $70-80K at least. And this doesn't even account for the development work Linux Mint put in on their side, to keep it as a desktop for their lineup.

Pantheon's Libgranite: Used almost exclusively in Elementary OS, they have had to build their own UI library for GTK in the wake of Libadwaita. This likely brought in at least 1-200 hours of labor, depending on how extensive the debugging and divergence is. $2500-$5000 worth of work.

CAFE: A fork of MATE 1.25, accompanied with a fork of GTK3 named CTK. All exclusively maintained by Zenwalker, for at least 4 years now. Assuming he puts in ~100-200 hours a year on this, mutliplied by the length of the project's life span so far, comes out to anywhere from $10-20k

LXQt: Arguably the earliest of the crew to see the writing on the wall, long before Budgie or Unity, LXQt was the end result of PCMan realizing that GTK was going down an unstable path. Now, to be fair, Razor was going on even before the merge, but the fact remains, this was labor. Since Razor's inception, plus PCMan's labor to port PCManFM, has to be at least 1000 man hours by this point, if not much more. $25k at least.

XAPP/Libadapta: Both maintained by Linux Mint, these projects were born of the goodwill that was burned by GNOME with the introduction of Libadwaita. Before this, GNOME apps were frequently re-used on other desktops, namely XFCE and LXDE. After GTK3 and Libadwaita later, this goodwill bit the dust, making a project to build cross-desktop base apps necessary. At least 1000-2000 man hours have been spent on this, combined, if not much more. Anywhere from $25-$50k.

Every single one of these projects requires: Hosting, Continuous Integration, local compiling to actually test the code beforehand, debugging, and more. Large amounts of Time, Electricity, and Money spent on all of these projects because GNOME did not listen to its user base. All of this waste, because of arrogance.

And the total cost of this now sits at upwards of $528,000 worth wasted. Over half a million dollars. And this is a *conservative* estimate.

The whole reason I am bringing this up, is because of the fact that a TDE-like approach, like what g4sra mentioned, is humanity's only shot at sustainable computing anymore. Get Hot New Stuff and Move Fast and Break Things have been a financial and ecological leech upon both mankind and the planet. Now, to add insult to injury, instead of making more durable hardware as we reach the limit of moore's law, we simply get AI copilots that nobody asked for, baked into the hardware, and made a requirement to force replacement. Even a halfway decent mid-tier PC costs at least $600 now, and falls apart in two years or less.

Last edited by UnixMan1230 (2025-08-18 15:22:29)

Altoid · 2025-08-18 17:41:45

Hello:

UnixMan1230 wrote:

Every single one of these projects requires ...

Very interesting post, never thought about it in monetary terms.

With more or less the same idea in mind, I've been posting about that for a few years now and the usual reply is "if you don't like it, why don't you blah, blah blah ...", obviously missing the point by millions of miles.

TL;DR

It seems that it's not possible (and probably won't ever be) to harness the creative work of the countless thousands of programmers, developers and contributors dedicated to this fantastic project started by Linus Torvalds in a way that that they just work together and come up with just 'one' 100% functional and scalable 'distro' instead of generating gadzillionz different flavours, forks of whatever is available out there.
One of the (few) negatives of open source and free licensing is that forking projects just for the sake of having their own project - "look Ma! I rolled my own distro!!!!" - has almost become the norm.
It is highly counter-productive and there is no logical reason for it.

Best,

A.

The officially official Devuan Forum!

#1 2025-08-15 13:37:11

excalibur dist-upgrade?

#2 2025-08-15 15:04:18

Re: excalibur dist-upgrade?

#3 2025-08-15 15:26:32

Re: excalibur dist-upgrade?

#4 2025-08-15 16:55:10

Re: excalibur dist-upgrade?

#5 2025-08-15 17:05:17

Re: excalibur dist-upgrade?

#6 2025-08-15 17:15:29

Re: excalibur dist-upgrade?

#7 2025-08-15 17:17:35

Re: excalibur dist-upgrade?

#8 2025-08-15 23:09:37

Re: excalibur dist-upgrade?

#9 2025-08-16 09:25:31

Re: excalibur dist-upgrade?

#10 2025-08-17 21:21:27

Re: excalibur dist-upgrade?

#11 2025-08-18 11:45:26

Re: excalibur dist-upgrade?

#12 2025-08-18 15:10:25

Re: excalibur dist-upgrade?

#13 2025-08-18 17:41:45

Re: excalibur dist-upgrade?

Board footer