The officially official Devuan Forum!

You are not logged in.

#1 Re: Other Issues » How does LTSP work ? » 2024-04-04 06:01:41

I just realised that devuan has ltsp5 . Not the newest ltsp incarnation. 

Announcing LTSP 19.08   
LTSP has been redesigned and rewritten from scratch as part of a GSoC 2019 project,
The new LTSP comes with a new goal:
Linux Terminal Server Project helps in netbooting LAN clients from a single template installation that resides in a virtual machine image or a chroot on the LTSP server, or the server root (/, chrootless). This way maintaining tens or hundreds of diskless clients is as easy as maintaining a single PC.

I.e. the focus now is in ease of maintenance, not in recycling old hardware. New technologies like diskless fat clients with UEFI, Wayland etc are supported, while thin client support is now reduced to "remote desktop with xfreerdp / x2go / VNC".

The old LTSP will now be called LTSP5 (it was first released in 2005), to distinguish it from the new versions that follow the year.month numbering.

..and i may add that newest ltsp has a systemd depedence.

A link to ltsp site would create confusion since the installation guide there refers to the newest ltsp.

So should i try to use a possible unsupported version or maybe try x2go ?

#2 Re: Other Issues » How does LTSP work ? » 2024-04-02 19:46:42

isnt that strange? Unix is multiuser. Unix uses simple tools to do the job. What do i miss ?

#3 Other Issues » How does LTSP work ? » 2024-04-02 05:53:44

Replies: 5

------------ update ------------
Devuan uses LTSP5 (not the ltsp19 from 2019 onwards) .(see related announcment of that change).

For documentation for newbies ,like me, i think the best entry point so far that i have searched is :

Devuan wiki / LTSP5

Also since ltsp5 is a devuan stable package and not in debian stable see:
devuan git page of ltsp5


I try to understand how LTSP works and ideally make it work in my Devuan system and it would be also cool to test devuan thin clients with qemu. Ideally i would like to administer a couple of home devuan pcs from my computer minimizing the fuzz to must admin each of them individually.
So i would prefer to install software in one system and the network-clients to access a unified root filesystem shared by all users but of course with access restriction do each user.

I found in the internet a distinction between ltsp with thin cilents and ltsp with fat clients(1).

It seems that an LTSP server offers (creates certain types) of images of installed systems to clients . And a client can use ipxe to boot with those images.

And a client while running an image (lets say Devuan stable) will be used AS a standalone pc with the difference that the filesystem resides in the LTSP server .

If my description covers correct the casual way ltsp is used today then the server is not a unix server for user in a traditional sense with each user having a home folder under a common root fs and sharing all the server's installed software but each client is practically another root filesystem installation with  the ltsp server acting as a file server ?

And then each user of a fat client must be his own sysadmin ? I mean , the LTSP server's sysadm is responsible only for setting a central server for distro images and the 'file server' ?

I guess a crucial question that helps to differentiate various setups it what are the responsibilities of a user ? Do we have one sysadm and many users or do we have one sysadm and many other 'sysadms'.
(1)By fatclient i refer to a pc that has cpu and main memory and can run programs but it does not necessarily has a secondary storage.

(a) How to create a Ubuntu 12.04 x64 LTSP server with 32bit thin clients  // 2011 LTSP5 tutorial .

#5 Re: Desktop and Multimedia » A devuan wiki page on virt-manager » 2024-03-17 05:30:36

I started using virt-manager and it kind felt ok . But it has that irritating config setup.
I wouldnt have problem trying command line. A small push with a link to a good tutorial
would help. I dont mean fancy one in a website that welcomes you with a thousand cookies
harvester companies. A simple one laying out simple concepts would be ok.

#6 Re: Desktop and Multimedia » A devuan wiki page on virt-manager » 2024-03-16 06:21:06

stopAI thanks for the suggestion. Searching recursively in virt-manager dependencies  ... ->

$ apt-cache depends --recurse --no-suggests --no-conflicts --no-breaks --no-replaces --no-enhances virt-manager | grep "^\w" | sort -u | grep 'virtinst\|libvirt-clients\|libvirt-daemon\|qemu-system-misc\|bridge'

...-> i see that  bridge-utlis  is the only that is not pulled.

#8 Desktop and Multimedia » A devuan wiki page on virt-manager » 2024-03-15 09:24:26

Replies: 7

---- UPDATE ---
I stopped the wiki page on virt-manager and instead i started a new one
on using Qemu in CommandLineInterface. Using QEMU from a CLI
---- endupdate ---

I started a wiki page on virt-manager in the

It's in the sandbox in .

I occasionaly use virt-manager and in now as in the past i had again difficulties in setting
it right in devuan. So i think a wiki page could function as focus in order to collect
instructions and tips to smooth things out.

devuan wiki needs i guess some thinking about how to organize things so for
time being i put it in the sandbox.

Also for now i choosed to  use as name the package name that i used for installation.

Any corrections and suggestions are welcomed .

#9 Re: Desktop and Multimedia » virt-manager works only when started as root. » 2024-03-15 06:36:15

Sorry. I guess you are right in guiding towards a pemission issue because eventually i had to change permission
related libvirtd options:

$ sudo nano /etc/libvirt/libvirtd.conf

unix_sock_group = "libvirt"
unix_sock_rw_perms = "0770"

$ sudo nano /etc/libvirt/qemu.con
user = "chomwitt"

group = "chomwitt"

And it was essential  to do also:

Upon opening Virt-Manager, it will default to the system variant (root) of the QEMU connection.
This can be changed to the user connection by going to: File > Add Connection.
Now select QEMU/KVM User session as the Hypervisor and click OK

#10 Re: Desktop and Multimedia » virt-manager works only when started as root. » 2024-03-14 23:21:49

GlennW thanks for the ideas but i think virtualbox is a different virtualization program .

For time being i followed the  instructions and it worked.

#11 Desktop and Multimedia » virt-manager works only when started as root. » 2024-03-14 20:45:32

Replies: 6

$ dpkg -l virt-manager
ii  virt-manager   1:4.1.0-2    all          desktop application for managing virtual machines

I have added myself to the libvirt group (to pass another error : unable to connect to libvirt)
but now starting virt-manager as a regular user i see 'QEMU/KVM - Not connected' and left clicking on the create a new virtual machine
i see a error window "Error : No active connection to install on."

But $ sudo virt-manager doesnt display that issue.

#12 Re: Desktop and Multimedia » Help needed with StumpWM » 2024-02-05 08:31:01

That is the devuan and the debian package stumpwm (2:22.11-3)

I just noticed that deb packages  have a package for ALL architectures. But arch has a binary for amd64.
So that explains my issue but not entirely. Because could it be that the lisp code is older and not 22.11-3 ?

#13 Re: Desktop and Multimedia » Help needed with StumpWM » 2024-02-05 07:19:42

@boughtonp  that is usefull info on version format.
But why installing in Daedalus stumpwm 2:22.11-3 when i run it says in the NEWS changes up to 0.9.6 version  and while running reports "version 1.0.1 compiled in 05 February 2024: (date i last tried ! ) ?

I dont compile it. that what it says. I just installed it.

And that happens even if i have sbcl installed not from the devuan repos but from the sblc site (newest stable version).

I also checked the arch stumpwm-22.11-3-x86_64.pkg.tar.zst .
It doesnt contain any NEWS .

Both arch and devuan-debian package contain a README that states :

Note: The recommended way to install SBCL is by downloading one of their
pre-built binaries available in their [web page][sbcl-platform-table] or build
it from source. Please do _not_ install SBCL using your distributions package
manager, especially Ubuntu. If you do so it is likely that you'll run into
problems when building StumpWM due to using obsolete versions of the

Also the debs contain a lot of lisp files. The arch package contains only the binary.

And in the deb the binary /usr/bin/stumpwm is a shell script:

sbcl --load "/usr/lib/stumpwm/load.lisp"

#14 Desktop and Multimedia » Help needed with StumpWM » 2024-02-03 15:42:27

Replies: 3

$ sudo apt-cache show stumpwm
Package: stumpwm
Version: 2:22.11-3

* Changes since 0.9.6
** in float mode windows can be resized with the middle mouse button

And while StumpWM is running Prefix + v  gives :  1.0.1 compiled at January 2024.

So what is the 2:22.11-3 for ?
Also the version i use has various bugs . If anyone has managed to compile (or wants to) the latest
version please tell me to help each other.

-------------------------------- my compilation effort
$ dpkg -L stumpwm
/usr/bin/stumpwm        // that is the stumpwm binary from the package Version: 2:22.11-3 in Daedalus.

Following the instructions more closely in the github page i managed to compiled.
I think it crucial to install a linux sbcl binary from the sbcl project page as instructed
and follow the instructions from there.

After installing the sbcl binary and loading to it the libraries needed we
compile the stumpwm src :
$ git clone
$ ./
$ ./configure
$ make
$ sudo make install

'/usr/local/bin/stumpwm'   is the directory that will be installed if we compile the github sources.
We put that path in .xinitrc and we are ok


Of course that leaves a bitter taste in my Devuan admin mouth. The package 'stumpwm' can not work with sblc  devuan package and that result in  an older version running?  And thus instead of 22.11-3 we get 1.0.1 ?

StumpWM main page
Debian tracker / StumpWM

#15 Re: Off-topic » [SOLVED] What is your view on major distros dropping X11 and requiring Wayland? » 2024-01-25 13:20:37

I wouldn't be so hasty to focus my attention to synthetic AI-driven bio-engineering. I dont trust the extremely well oiled (with billions) hype engines that trigger massive diversion of public or shareholders money and value to exciting new techs that instigate FOMO feelings to the public and to the investors while more pressing ecological and society problems get under the radar . On the contrary that attention diversion highlights  the great power imbalance and pathological power concentration on the mass media , and mass social platforms.

Can we really entrust our future to companies that build their fortunes on parent's money for pc room-cell gaming? Last time i checked farmers in Taiwan were suffering in order to sustain the water-hungry TSMC. Can we trust steam to guide the future of the display server??

That's one reason that i prefer more often to look back . Like in How X started..

#16 Re: Off-topic » [SOLVED] What is your view on major distros dropping X11 and requiring Wayland? » 2024-01-22 18:58:31

Touche!  big_smile

Seems I hit a nerve. FWIW . . . the rage quit and deletion was completely unnecessary and deprives the forum of your thoughts which had value historically if not practically.

Are you answering ? Or fueling with more facts the nonpoliteness that i attributed to you ?

#17 Re: Off-topic » [SOLVED] What is your view on major distros dropping X11 and requiring Wayland? » 2024-01-22 17:11:37

Oh is that all! So easy to say. So useless to say. Show us the fork YOU are going to initiate and we will take your babblings seriously. Hot air does nothing but add to our collective climate woes.  wink

I hope you can make sense of what you write. Because it seems to me that your response except from lacking politeness is also so easy to say  and so contradictory to say.
My 'babblings' will get elevated to not 'babbling' status when i present in front of you code ?
Show me your code ? That is the babbling status elevator ?  Really ?  That old opensource mentality ?
Have you read my arguments? Who drives Wayland developemt ? How  386 arch opensourced X11 started?
Have you tried to answer 'Why no X11 fork exists?'  And lastly have you pondered how ethical are the gaming money?

And i guess all other comments on an offtopic thread have pass you babbling status filter. 
Even comments that promote possibly false inflammatory dichotomies X vs Wayland .

I guess you intend to program the heck out a harmonious libre community  since last time i checked politeness
hasnt been incorporated in any existing programming language.

#19 Re: Off-topic » [SOLVED] What is your view on major distros dropping X11 and requiring Wayland? » 2024-01-21 09:24:34

I'd like to remind my relative posts ;
How X started
Copy paste , a hard problem?

A very short recap of my current understanding is that historically speaking a Display Server (DS) is a concept broader than a rendering frame perfect protocol. And than Wayland can be seen as an improvement in the a DS's RP (Rendering Protocol) and not as a replacement of X11. Why ?
Because X11 was a DS with network-centric nature from day one.  So i think posing Wayland as X11 replacement creates confusion and harm.

For those interested in detailed arguments in favor of my claim you could read my posts where is the results of days digging in the web for X and x386 history.

As for the future..
I am pessimistic too. But not only of X . I am pessimistic of Steam and Codeweavers too . PC - roomcell gaming is a selfdefeating aim. Contradictory .I am also pessimistic of the perfect frame goal.As Wittgenstein would said every construction is a short of a prison. And i guess prisons, how matter fanciful ,after a while they start to suck more and not suck less.

Game money and gpu money is  currently in the driving seat. That could end . I think ,  as in the X case ,  the 'new' wont came from the hackers . They look blind sometimes. It will come from society new needs. BUT i think in order for libreIT world to respond quiclly to those needs X11 networking and platform - vendor independence spirit must be kept alive.

Lastly , i think it very basic discussing about X and Wayland to answer why there is not a fork ? Isnt that supposedly the spirit of libresoftware that shines ?  Could it be that Wayland is a fork ? And then the case is simple. One group of people must step forward and  do another fork focusing not on the rendering site but on the network centric side. And that means proposing a common standard - protocol for remote - desktop , remote graphics . And remote-network-graphics cant be frame perfect. They must be typed and semantically sound. A pixel or mime cant be a sound type for that kind of remote computing, And you cant have either one type-system. You must be able to accommodate many.

#20 Re: Off-topic » waylandows is crap » 2024-01-12 20:25:34

2D drawing primitives that almost nobody uses because these days most programs render directly to buffer instead, or use the GPU's 3D rendering functions.

You mean that certain set of 2d primitives is outdated or the idea of a network centric DS having 2D primitives is outdated?

#21 Re: Off-topic » Search babble » 2024-01-12 16:00:32

I think indeed an ethics purist could vanish form nigthmares because before even purchasing a motherboard s(he) must eat. And i  guess many we are still eating meat and i do believe that animals could have consciousness. Let's not start on what entities funded the internet..

So i prefer practical ethics. Meaning , that in a community with certain objectives if alternatives do exist and have been weighted and there is a consensus that choiceA promotes those objectives then the community should encourage by various means that choice .

So in that path of practical ethics in IT space , i think that there are alternatives to google . I try to use searx instances and in my kid's pc runs only Devuan + netsurf with wikipedia as search engine. Ofcourse i am not sure as to how much a metasearch that uses google search can be valued as an alternative or as a facelift. Anyway even more i value the fediverse alternatives.

But speaking of google.. i think the surface is way to more slippery that we can imagine .

Many libre-land organizations and projects have been funded by it. Even FSF has dipped it's fingers in the sponsor's honey pot. And then.. if we start that discussion i think it unavoidable to discuss the donation's ethics . And my opinion is that on that issue the surface has been more than slippery.

So slippery that a couple of years ago almost all know libre-land organizations that had been practising happy sponsor hunting from google, microsoft , redhat etc were willing to hump on on the RMS lynch bandwagon  in a witch-hunt that acted also as a much needed distraction of the funding scandals from eipstein that had shaken from the ground the MIT!..  So ethic wise the libreland's most prominent organizations seems to me in such a bad shape that they are in danger to be tagged as a Grace Field House in the Promised Neverland manga..

How else to call organizations that under a freesoftware flag promote google hacker initiation events. (now this is getting also political!!..because the power that google has amassed makes it eventually a political entity..and politics are too much heated those days...).Or how else to call organizations that under a libreland flag call has github as sponsor , lay roses for programmers to use that forge and then we realize that the AI agenda needed fresh ideas to feed upon..

As to the hardware i think projects like MNT reform show the way forward and open risc architectures seem a good alternative too.

And last but by far not least.. we should discuss also the gaming ethics. Empires,fortunes and mega-egos  have being build on gaming , and even nvidia that promotes itself in the well-oiled hype engines as futuristic AI -cloud company , a solid argument could be made that the kid's parent's purse have being the humble (but ethicwise questionable) fuel of those mega-ambitions.  I think gaming is the elephant in the closet that nobody dares to touch and surely that closet has monsters in various forms.

#22 Re: Off-topic » waylandows is crap » 2024-01-12 13:19:49

I think minimalist can mean not only the code size but also the graphical elements . Both meanings have be mentioned in that thread. I dont pay much attention to the code size because in my case it doesnt make a difference. On the other hand if a wm in the spirit of X offers maleability and less policy,that could help in creating more interesting workflows..

Speaking of workflows, a supervisor is needed to manage automated workflows. For those interested, the user-workflow supervisor idea 's theoretical exploration (user runit) has been discussed in the elist.

In my case interesting workflows would need support of multi-tagging of windows among other things.

Also speaking of maleability ,X allowed me to pass  in a couple of minutes from i3 to testing StumpVM.
So 1983's X DisplayServer project legacy seems to me to favor recombinations . I wonder , evolutionary speaking , in software terms, if Wayland's more integrated style that binds a DS's rendering protocal (DS's RP) and a certain  WM  and call that DS is going to be fruitfull in the long run..

I say DS's RP because as i have proposed in another thread Wayland is a project that doesnt align with X's 1983 legacy , philosophy, incubator enviroment ,and generaly it's dna. That misalignment is painted in the most striking colors in some presentations in the www that present X as a project that has been stagnated in a bad IPC layer. But i propose a different intepretation. That instead ,the Wayland project represents , if not a stagnation, a narrowing in the scope of a network-centric DS that the Athena X project represented.

Wayland is a notable effort to improve the rendering protocol but by various circumstances of fate , it promotes it self , falsely, as successor to X11 . But if you study the project history, easily, we can see that x386 - xfree86 and xorg represent groups focused only on one aspect of the DS protocol stack.

Thus i think it's more fair to say that Wayland it's an extension or a DS's RP fork. I would like to being able to use Wayland compositors but as co-operating components - extensions of a X12 or Y project that promotes and improves while at the same time aligns and respects the initial unix-centric network-centric philosophy of Athena's X.

#23 Re: Off-topic » waylandows is crap » 2024-01-12 00:29:25

I try last days stumpwm. I think it was inspired by ratpoison but they wanted it to run on common lisp.
It certainly seems minimal as quickfur describes.

#24 Re: Off-topic » Copy paste... a hard problem ? » 2024-01-07 12:41:02

quickfur doesn't your argument applies to every aspect of a display protocol ?

if wayland specifies a protocol for committing pixel perfect frames, it would still be up to the relevant applications to implement it in a sane way before it would work. If one of the authors is lazy and decides to just ignore it, or implement it minimally  then you'd get various screen artifacts

A protocol is a contract . To be of any use the programmers must study it and  use it.  A lazy programmer would use could use it wrongly.
So i think we could make a DS protocol with perfect copy paste. And it would be a perfect copy paste desktop for those apps that conform to the protocol rules.

Now we are in a situation that i hear horror stories of a user that must know what apps run on xwayland and what not in order to predict if copy paste or drag and drop would work.

So we have a pixel perfect desktop BUT it seems that 'we' miss something essential . And that essential could be a problem not amenable to hackers skills but out of their reach. That's an idea that i explored for days in my post How X started.. .

I mean the speed by which X was used and prospered is in great contrast with what is happening with Wayland. Wayland is not X. Shouldnt be seen as a display server protocol DSP replacing X but either as a pixel perfect extension or as another type of display protocol aiming perfecting a certain aspect of the a display protocol.

I personaly think wayland is not a DSP. Wayland is a DSP's  perfect frame component.

All those companies and institutions that are in the remote desktop 'business' i  think would agree. They all implement various free or proprietary protocol to share desktop rendered data across various networks. (the initial X use case) . So those are protocol controlling display data sharing. There we are.
Display Server Protocols chassing a differerent rabit hole!.
A DSP must encompass more.  For me a generic DSP needs a RenderingProtocol (RP)  But the RP  cant be the DSP .

So i think 'we' the IBM-PC clone era architecture users are getting dangerously close to the 80's !!.
Back then we did have again x86 PCs. And we used various 'display servers' .  But 1) they were proprietary 2).  they were aligned closely to a pc desktop not-networked view of computing .

ps: I just realized that another issue i have with firefox in sway could be also related to copy-paste problem. I cant drag and drop links inside bookmark menus . Something that i was doing the last years . On top of that i cant copy paste text to the 'foot' wayland terminal from many of my favorite apps.

#25 Re: Off-topic » Copy paste... a hard problem ? » 2024-01-06 11:40:02

I'm not blaming.  I just say that although you have a difficult issue you prioritize the perfect frame .

Isnt it essential, to productive desktop workflows, to have desktop apps exchange data ?

In my devuan+sway installation indeed i dont see graphically speaking anything buggy but i cant copy paste simple text between some apps breaking my workflow.

And isnt it the essence of a DS protocol to laydown some rules for the clients . And the clients should coooperate in order to use the DS's services and resources ? 

So my question is could a protocol BOTH apply rules to committing pixel perfect frames AND enforcing data exchange protocols ?

An tty-analog example could be Unix. To have pipes in Unix a unix pipe-friendly program must follow some conventions. Use stdout by default , use ascii. If a program decides to write to the console not in ascii but in a finer-grained private encoding then doesnt that brake the pipe ?

Also isnf a legitimate display server functionality to enforce certain type-constraints on data written to it's graphical memory so that sharing could be easily exposed to it's client as a service? (generally type constraints have also a security bonus).

Shouldn't also the integration of  the window manager in the display-server compositor made the matter easier for wayland (if there isnt not any protocol inherited constraints) ?

In X11 you have 4 parties involved: X11,windows manager, target app, source app

Board footer

Forum Software