KLV-Airedale-alpha Released for Experimentation and Improvement by the Puppy community

[wiak]

So to test octoxbps, should just need two simple steps.

1. Edit the text config file located at /etc/sv/agetty-autologin-tty1/conf and replace root with weedog
You can do the above from Rox or Thunar - just find that file and right click open with geany.
(or, in a terminal, enter command: geany /etc/sv/agetty-autologin-tty1/conf)

2. Now switch to weedog desktop via start menu:

Applications -> Log Out

press the big Logout button, which takes you to a terminal, and at the terminal simply type: exit
If you get asked for a password for weedog (can't remember if you do), I think it might simply be weedog.

Then your previously installed octoxbps should work allowing you to check it out.

Reverse above procedure to return to being logged in as user root. Em... except, once you have successfully reached desktop as user weedog say, you'd have to start filemanager as root user via (from a terminal commandline window):

sudo rox
or
sudo thunar
or simply use:
sudo geany /etc/sv/agetty-autologin-tty1/conf

prior to changing name weedog back to root; otherwise it won't save since weedog user doesn't have permissions to save into /etc area.

Just try it. It is simpler to do than to explain it. Just that one quick text edit involved.

Could likely be easily automated using sed in a script so from user point of view switch-user made just into another Start Menu (Applications) choice.

[wiak]

Obviously everyone knows that Thunar: Edit -> Preferences: Behavior tab allows you to select single click for activating, but I'll say it anyway just in case anyone curses double-click need. That could probably be arranged in the distro build plugin during the root filesystem creation (but sometimes it is not so easy to automate such choices at that time).

[wiak]

By the way, apart from that fancy octoxbps, there is a nice commandline helper app for xbps available. I don't use it, but it does help avoid learning all the many xbps option tricks:

xbps-install -S vpm

It's github page is: https://github.com/netzverweigerer/vpm

I don't know if it covers full xbps functionality; it might... I just don't know. It works fine as user root (or start with sudo if logged in as normal user).

vpm --help-pager (for paged help, or use vpm | less)

vpm deps firefox (shows dependencies of package firefox)

and so on...

The great thing I see about vpm though, is that it shows you the xbps command that it uses underneath, so a great teaching/learning-xbps tool, and it is tiny to install. And would be pretty simple to write a bash/gtkdialog or yad frontend for vpm.

[retiredt00]

KLV is a nice fresh idea and for an early alpha runs ok.
Α front end for wiakwifi (or another more user-friendly tool), a fix for octoxbps as suggested above and presence OOTB, mounting things in one place instead of /media /mnt and /run (to minimize hunting - check TinyCore for howto), using one WM for everything and facilitating use of pet packages (this is the puppy forum after all ) will further improve first impressions.
But then ιt would be a beta

However, I admit using puppy after some time I was totally confused with KLV as it does not appear to be a derivative of any puppy. More important, does not behave like any puppy or has the familiar “puppy apps” and “puppy look”. Is rather a WDL-void “upgrade” as far as I can tell looking briefly around here. This might alienate some users that may come with different (puppy oriented) expectations for a "puppy derivative" (I'm just persistent at times ).
I would think that “dog incubator” might be a more suggestive place at this stage.
If/when puppyfication improves or this new look and functionality is puppy-adopted, move it into “official puppies”!

In any case, thank you for this interesting mutt

[wiak]

I was totally confused with KLV as it does not appear to be a derivative of any puppy.

No it is not derived from Puppy. It aims to becoming a hybrid using useful components from any forum discussed/developed projects and distros. Seems to be a fair amount of input coming from Puppy users and also other Dog users so the distro becomes whatever it becomes. I guess that's why Kennel in title. Also it's not thought of as a one off; may likely come other different style Kennel Linux distros. No restriction on which component selection. Final designs are up to the whole forum community. As you probably know there have been various distros worked on in this forum by its members for over 8 years now. The members are the forum.

Whilst location doesn't matter very much at all, since it is found easily enough, it might be good to rename "Dog incubator" to "Kennel incubator" - your guess is as good as mine.

[wiak]

mounting things in one place instead of /media /mnt and /run

In general, per standard UNIX/Linux practice, drive partitions are generally found under /mnt.

I'm not in KLV-airedale at this moment, so can't check, but I think you will find most sfs addons (the layers) individually under same /mnt/layers directory (sometimes under RAM in there). You can see their internal contents by looking there.

Up to rockedge how else he wants to configure final mounts based on community feedback - I guess he could assign /mnt/home in Puppy kind of way via a mount bind or symlink situation. For Kennel Linux using any WDL components I'm not planning to change how the components work, but I'm happy maintaining these pieces, and helping out sometimes when I can overall. Certainly, I think rockedge experience with Void xbps packaging and runit comes from WDL_Void work; same for me of course, though after starting WDL design with Void repo use I myself concentrated on Arch Linux build to the extent that I am very rusty about Void runit and xbps now (though reading up on that again) - rockedge certainly has far more experience building with Void Linux packages now than I do, but he also uses Puppy builds a lot, so knows how to configure JWM and so on in Puppy way so his choice of build always reflects that experience I think.

[rockedge]

My new optical coupler failed! No Internet until December 20. Only have phone and limited data. I will keep reading and experimenting, but can't type too much!

[fredx181]

How to create working Startup folder (in /root):

- Copy attached "start-up" script to /usr/local/bin (remove fake .gz and make executable)
- Create Startup folder mkdir -p /root/Startup
- Edit /root/.xinitrc, find the line 'exec xfce4-session', and before that add '/usr/local/bin/start-up &', so will look as this:

Code: Select all

/usr/local/bin/start-up &
exec xfce4-session

Add the scripts (or symlinks) that you want to start at login to the Startup folder and reboot (or logout and back in).
For example, I added a symlink to wiakwifi (after configuring) so that my ethernet connection starts automatically.

[rockedge]

Nice work fredx181. I can't try yet, but will the script allow dropping to console like when removing the "exec" from the.initrc? I really like using the /root/Startup directory like in Puppy

[fredx181]

will the script allow dropping to console like when removing the "exec" from the.initrc?

Sorry, don't understand what you mean exactly.
I tried removing the "exec" from "exec xfce4-session", logged out and got a few xterm windows (I have xterm installed) (and xclock), typing startx in a xterm window didn't bring me back to the Desktop, btw.
EDIT: Without having xterm installed I tested, it makes no difference when removing exec from "exec xfce4-session", logging out and back in with startx, the Desktop appears.

[keniv]

@fredx181

How to create working Startup folder (in /root):

- Copy attached "start-up" script to /usr/local/bin (remove fake .gz and make executable)
- Create Startup folder mkdir /root/Startup
- Edit /root/.xinitrc, find the line 'exec xfce4-session', and before that add '/usr/local/bin/start-up &'

I've followed your instructions above but have got stuck at the instruction below.

Add the scripts (or symlinks) that you want to start at login to the Startup folder and reboot (or logout and back in).
For example, I added a symlink to wiakwifi (after configuring) so that my ethernet connection starts automatically.

I can't find wiakwifi to make a symlink. Normally I would use pfind to find wiakwifi. Can you point me in the right direction as far as the whereabouts of wiakwifi is concerned.

Regards,

Ken.

[fredx181]

Hi Ken, wiakwifi is in /usr/local/bin.
So to create symlink in Startup, in terminal:
cd /root/Startup; ln -s /usr/local/bin/wiakwifi wiakwifi

[keniv]

@fredx181
Hi Fred,
Thanks for that. However when I rebooted wiakwifi did not start. I did check that the symlink was in /root/Startup and it was. I did wonder about what I thought might be an extra, second "wiakwifi" in your code (please see below).

Code: Select all

cd /root/Startup; ln -s /usr/local/bin/wiakwifi wiakwifi

.
I tried making the symlink both with and without the second "wiakwifi" but neither worked.

Regards,

Ken.

[keniv]

Hello All,
I have mentioned what I think is a lack of a firewall in this OS in a couple of posts. Running without a firewall is worrying me. I wonder what others on this thread are doing with respect to using a firewall. Should I be concerned about this? Does the OS contain a firewall that I have not found? If so can somebody tell me how to set it up? If not can anybody tell me how to download and install one?

Regards,

Ken.

[fredx181]

@fredx181
Hi Fred,
Thanks for that. However when I rebooted wiakwifi did not start. I did check that the symlink was in /root/Startup and it was. I did wonder about what I thought might be an extra, second "wiakwifi" in your code (please see below).

Code: Select all

cd /root/Startup; ln -s /usr/local/bin/wiakwifi wiakwifi

.
I tried making the symlink both with and without the second "wiakwifi" but neither worked.

Regards,

Ken.

Hi Ken, the code should be correct, the "second" is the "target" (to create in /root/Startup) so not sure what's wrong at your end.
You did configure wiakwifi earlier ? I configured it to use eth0, not sure if it works as above (symlink in Startup) when configured to connect through wi-fi.
EDIT: You may want to try adding another script in /root/Startup to see if the startup mechanism works anyway for you.

[rockedge]

Working on different firewall approaches. Can not do much at the moment, Internet is down and will be repaired, earliest this Friday.

But I have begun looking into this. Using the NetworkManager iptables is included which can be configured. Most firewalls are frontend GUI's for iptables

There's lots of room for development here concerning a firewall

[mikewalsh]

@fredx181 :-

Ooh, you little beauty, Fred! D'you know, I hadn't realised until this last few days just how much I rely on /root/Startup for so many things. Chrome starts up in the background via --silent-launch, Thunderbird fires-up in the background using BirdTray (Qt5-based replacement for FireTray); gKrellM starts up, etc, etc. Mind you, it appears I was successful in configuring gKrellM to start-up via a runit 'daemon'; I thought I'd cocked it up last night, but when I fired-up "Airedale" earlier on there it was on the desktop straight-away.

It's a lot of messing about, though. Will will probably accuse me of being lazy ( ) , but using /root/Startup is SO much easier. I haven't even got around to looking at NetworkManager properly, yet; the one time I tried using it to set up a connection, clicking on it did absolutely nothing, so I've stuck with "wiakwifi". Does what I want, so like you I've now sym-linked it into /root/Startup. Fires up a treat.....

Thanks, buddy!

------------------------------

@keniv :-

Ken, there IS gufw in xbps; it's the GUI 'front-end' for 'ufw' (Uncomplicated Fire Wall). I've installed it OK, but it's not showing up for some reason.

Anyone got any thoughts on this one?

Terminal is showing me this:-

Code: Select all

bash-5.1# gufw
Unable to init server: Could not connect: Connection refused
Unable to init server: Could not connect: Connection refused

.....and then this:-

Code: Select all

(gufw.py:9049): Gtk-CRITICAL **: 20:25:00.688: _gtk_css_lookup_resolve: assertion '(((__extension__ ({ GTypeInstance *__inst = (GTypeInstance*) ((provider)); GType __t = ((_gtk_style_provider_private_get_type ())); gboolean __r; if (!__inst) __r = (0); else if (__inst->g_class && __inst->g_class->g_type == __t) __r = (!(0)); else __r = g_type_check_instance_is_a (__inst, __t); __r; }))))' failed
/usr/lib/python3.10/site-packages/gufw/gufw/view/gufw.py:77: Warning: g_object_set_data_full: assertion 'G_IS_OBJECT (object)' failed
  self.builder.add_from_file('/usr/share/gufw/ui/gufw.ui')

(gufw.py:9049): Gtk-ERROR **: 20:25:00.690: Can't create a GtkStyleContext without a display connection
/sbin/gufw-pkexec: line 2:  9049 Trace/breakpoint trap   python3 /usr/lib/python3.10/site-packages/gufw/gufw.py $1
bash-5.1#

Looks to be summat to do wi' Python3. I HATE Python; you can't trouble-shoot it in the normal way, 'cos it's nowt but a bunch of scripts.....

(*grrrr....*)

Mike.

[wiak]

It's a lot of messing about, though. Will will probably accuse me of being lazy ( ) , but using /root/Startup is SO much easier.

Well it is 'lazy'. I know this become I am generally lazy too so I've nothing at all against alternative startup methods such as a /root/Startup folder. Only reason I say it is lazy is that many of us tend to go to generic methods such as that rather than bother becoming experts at the likes of creating and configuring runit services. Problem is, we never become masters of the likes of runit (that's why the word 'lazy' is appropriate) whereas we could have. I blame sysVinit for these habits cos that old init system is more a case of cobbled-together shell scripts, whereas runit /etc/sv services take a logical format that could be learned and mastered.

However... there is absolutely nothing wrong with both methodologies existing together /etc/sv type services (which benefit from being able to be controlled, status checked and so on by runit management commands such as 'sv') and startup folders such as /root/Startup folder. In fact, the startup folder itself could be run from a boot time runit sv, and/or some of the contents inside /root/Startup could be links or scripts that refer to runit /etc/sv services. Nevertheless it would be nice to reliably learn and understand runit service creation and handling though (and I say this to myself, who keeps only learning snippets of that).

I don't know anything at all about NetworkManager either by the way - I also went through the stage of just running it and getting nothing. However, I then simply started it as a service (per the Void Linux Documentation on starting services) and next time I booted it was just running and the key control mechanism was that wee network-manager-applet that was also installed - it just asked me for name of my WLAN and password and bang, it worked - I have no idea of how or the details but it always auto connects now. wiakwifi method is simplicity itself and I've previously used that method on other distros for a couple of years, but it is not so capable as NetworkManager in alternative network configurations, but okay...if connection drops it won't reconnect automatically (unless itself run as a service) for example. Actually, truth is, wpa_supplicant (for wifi connection) is run directly by wiakwifi, but Void Linux provides a runit service to run wpa_supplicant (and auto rerun it if it drops) so special Void-compatible version of wiakwifi could be written (but likely won't be...). Netmanager won't work if wiakwifi already running - since wpa_supplicant will be competing.

In summary, I have nothing against /root/Startup approach - it is very useful addition.

Personaly I never use a firewall, by the way - I probably should, but even with Puppy I generally don't check the firewall box since if I'm 'experimenting' I fear things won't work because of the firewall blocking protocols I'm trying to use... Lazy me again - I should master firewall configurations instead of avoiding them.

[wiak]

Ken, there IS gufw in xbps; it's the GUI 'front-end' for 'ufw' (Uncomplicated Fire Wall). I've installed it OK, but it's not showing up for some reason.

Anyone got any thoughts on this one?

I think you are on to a good firewall solution. It may be that before gufw can do its business you need to start a ufw runit service (see, back to that again...). Thereafter, it may all work. One thing that might break that process is that ufw apparently uses a firewall method called 'netfilter' rather than using 'iptables' - I read on Arch Wiki (about ufw): (EDIT: correction - actually I think netfilter uses iptables underneath so problem is more to do with not using ufw netfilter-based service and separate iptables service)

https://wiki.archlinux.org/title/Uncomplicated_Firewall
Install the ufw package.

Start and enable ufw.service to make it available at boot. Note that this will not work if iptables.service is also enabled (and same for its ipv6 counterpart).

I imagine similar needs to be done to get ufw (and thus gufw) working in Void Linux.

Note that I often refer to Arch Wiki no matter what Linux I'm actually using. It is just such a good Wiki. However, Arch uses systemd and Void Linux uses runit so you have to disregard the systemd parts and think 'runit services' methods instead.

https://wiki.voidlinux.org/Firewall_Configuration
[Note (wiak) that Void wiki is considered deprecated now, so this may be old info. For example to check ufw status 'might' now be:
sudo sv status ufw; use Void Documentation preferably when appropriate.]

Code: Select all

ufw - Uncomplicated Firewall

Basic firewall rules (deny incoming, allow outgoing) can be established by default by installing and enabling ufw:

 $ sudo xbps-install ufw
 $ sudo xbps-reconfigure ufw
 $ sudo ufw enable

To check whether the ufw firewall is active during the session:

$ sudo ufw status

Annoyingly, Void Linux no longer has dedicated forum but uses reddit for that purpose instead. I found the following at Void reddit area (NOTE: I didn't paste the following - whole text just appears when I put in the link so is a function of this phpBB setup; just click on the following for possible gufw reddit solution):

[mikewalsh]

@wiak :-

Actually, digging-around further it appears that gufw/ufw is an absolutely horrendous mish-mash of not only Python3 but GTK-dialog AND Glade (for good measure). Some of those errors look to be GTK-dialog-based, some Python-based.....

Personally, I wouldn't even know where to begin if it's a UI fault. You can set the backend firewall going.....and it'll then start at boot every time. Just enter

Code: Select all

ufw enable

Simple as that, and only needs entering the once. To check whether it's running or not, enter

Code: Select all

ufw status

(Years since I've used this thing, but it's gradually coming back to me....) And, yes, I have to agree about the Arch Wiki. Best source anywhere for just about ANY Linux-related documentation, that's for sure. I often use it myself for all sorts of odds & ends; if you can't find it there, you almost certainly won't find it anywhere else.

Mike.

[wiak]

@wiak :-

Actually, digging-around further it appears that gufw/ufw is an absolutely horrendous mish-mash of not only Python3 but GTK-dialog AND Glade (for good measure). Some of those errors look to be GTK-dialog-based, some Python-based.....

Python can be tricky (so many optional parts that can well be needed by some Python-using apps). But that's one of the good things about using an upstream official package manager, because 'usually' a whole team of people up there have carefully tested each app provided in the repo and the official package manager (xbps in this case) will correctly download all correct Python-needed dependencies. That's also the case with Arch Linux based distro that uses official pacman. In practice that works really well with pacman installs; however, whether it works so well with xbps installs depends on the diligence of the Void Linux package maintainer teams - if some dependency not recorded correctly by their gufw package maintainer then that is a problem - normally, such problem is best addressed by contacting Void team - my only gripe about Void being that they no longer maintain a great forum; relying as they do on reddit Void area, but that is not such a great way to officially contact Void maintainers I feel (maybe it is, but I guess they instead respond more definitely to Void packages git issue notifications, which is much more formal and thus not so accessible to normal users).

Void team is nothing like as big as Arch Linux team so alas we cannot expect their documentation to be up to Arch Wiki standard but usually their package maintenance in terms of Python deps or whatever is pretty good. Time will tell for particular cases.

[wiak]

What I find really attractive about Void Linux is not (quite) the perfectness of its package management (Debian apt/dpkg repos and Arch Linux pacman repos are supported by huge teams so probably better in that respect when used with these official package managers), but the fact Void Linux itself is not a huge enterprise but does compile its own packages. So it is similar to original Puppy in terms of its independence from the big boys, but also does a pretty good job at package dependency maintenance - it fits in very well thus into using its upstream developement along with overall distro philosophies embraced by distros on this forum. It also continues to quite a large extent to also support 32bit releases (and other architectures). So I find it a better fit than Debian/Ubuntu/Redhat to that independence from the big distros.

Arch Linux is pretty independent in outlook too though, which is why I also like it (and specially because it does have great documentation which applies perfectly when using pacman of course).

Void Linux strives for efficiency and ability to break down packages into smaller units, which may occasionally be a problem rather than a blessing, but user can easily create 'meta' larger-package-recipe-templates from these (and its xbps package manager is available as a standalone static compiled binary) - so very handy for building even minimum sized Puppy-like distros - but neither Void nor Arch themselves provide the kind of frugal save persistence efficient/flexible install/control and sfs-load-handling modes we are used to - otherwise we would maybe not be on this forum at all and just using the likes of Arch and Void official distro downloads for all of our wishes and needs... Since Puppy itself also moved to using upstream repos, it is simply logical to also consider using the official upstream package manager that was specially designed to reliably work with these upstream repos, which is part of what is being done in this particular Kennel Linux Void Airedale release. There is however nothing to stop a woof-CE Puppy builder adopting xbps into an officiial Puppy Void release, and I believe/know that will also be a (somewhat related) result/option of this current work and the jointly-useful discussions here.

Related comment on occasional Void Linux repo maintenance weaknesses:

[wiak]

One last thought about the gufw firewall matter. I guess it is only gufw that has Python as dependency. Nothing to stop someone writing a simple bash/gtkdialog/yad frontend for underlying ufw firewall app. Since ufw has been around for well over a decade, it might even be the case that one of Puppy's simple firewall programs uses ufw underneath or simply set up iptables more directly so could be made to work in KLV-airedale.

To complicate choices... there is also not nftables instead of using iptables.

https://wiki.archlinux.org/title/nftables

I read (somewhere) that ufw now supports either as a backend (and thus gufw will do the same). However, I know nothing about all this and I agree with rockedge that it is worthy of a lot of dev research work being done by users of KLV. I have noted the availability also of firewalld, that does provide a Qt-based tray applet (GTK3 having apparently depreciated support for some tray applet uses?):

https://firewalld.org/documentation/uti ... pplet.html

firewalld looks interesting, but maybe over-complicated? Newest versions all use nftables underneath.

https://firewalld.org/

Who is using it?

firewalld is used in the following Linux distributions as the default firewall management tool:

RHEL 7 and newer
CentOS 7 and newer
Fedora 18 and newer
SUSE 15 and newer
OpenSUSE 15 and newer
Available for several other distributions

firewalld does not depend on NetworkManager, but the use is recommended. If NetworkManager is not used, there are some limitations: firewalld will not get notified about network device renames.

Applications and libraries which support firewalld as a firewall management tool include:

NetworkManager
libvirt
podman
docker (iptables backend only)
fail2ban

[wiak]

Note:

irc for void linux questions/answers:
https://voidlinux.org/news/2021/05/libera.html

forum for void linux questions/answers:
See this old Void news link for all the info and gossip about this you need!
https://voidlinux.org/news/2018/11/forum-is-down.html
But, as Void Linux team say in top link, they moved IRC from freenode to Libera.

[rockedge]

@wiak I was thinking of just borrowing the FIrewall GUI from Puppy Linux. I also think it will work.

As soon as the Internet connection is repaired I will finish alpha5. This will include hopefully all of the pieces and ideas members are contributing.

I did experiments with a WDL-Void-jwm-rox that I added a lot of Puppy Linux pieces and structure too. It was easier adding the structures that Puppy expects to the Void Linux base structure. I did not take it as far as it could go but enough to see the potential. When a Puppy Linux Void or "VoidPup" and runs with xbps and runit it will be a big contender for sure.

[rockedge]

Sorry, don't understand what you mean exactly.

I changed /root/.initrc to this below, because on some systems I could not drop to console. It would logout loop back into the X server desktop I just logged out of unless I removed the exec

Code: Select all

#!/bin/sh

userresources=$HOME/.Xresources
usermodmap=$HOME/.Xmodmap
sysresources=/etc/X11/xinit/.Xresources
sysmodmap=/etc/X11/xinit/.Xmodmap

# merge in defaults and keymaps

if [ -f $sysresources ]; then

    xrdb -merge $sysresources

fi

if [ -f $sysmodmap ]; then
    xmodmap $sysmodmap
fi

if [ -f "$userresources" ]; then
    xrdb -merge "$userresources"

fi

if [ -f "$usermodmap" ]; then
    xmodmap "$usermodmap"
fi

# start some nice programs

if [ -d /etc/X11/xinit/xinitrc.d ] ; then
 for f in /etc/X11/xinit/xinitrc.d/?*.sh ; do
  [ -x "$f" ] && . "$f"
 done
 unset f
fi

xfce4-session

[bigpup]

I am just wondering what this OS is trying to be?

1. A Puppy version, using Void Linux, as the source of always needed core Linux files and programs?

2. A Void Linux version, that looks and works like Puppy Linux?

Sure seems like a lot of reinventing stuff is having to be done.
Making programs, that did work in Puppy Linux, now work in this OS.

I am not sure what to really do with this OS.
Rockedge or someone, what is it, 1 or 2?

Note:
I changed to being logged in as weedog.
Needed for running Octoxbps

Octoxbps works great.
Very simple GUI, but it does give visual listing of available stuff and indication if installed or not.
If you stick with xbps.
This should be how to run it.

However, now that I am not root.

Got to use sudo to even run wiakwifi to get network connection.

Got to use:

Code: Select all

sudo loadmodulegui

in terminal, to get SFS load GUI to run.

Overall, as weedog, a lot of programs seem easier to use.

Thurnar
Default applications seem to be recognized better, as the one to use for a specific file, you click on.

The drive icons on desktop
Seem to know what to do when clicked on.

I installed ntfs-3g package
So ntfs format can be accessed.

[bigpup]

Wallpaper:
.
.

Void__UI_Screenshot.jpg (66.94 KiB) Viewed 553 times

[wiak]

When a Puppy Linux Void or "VoidPup" and runs with xbps and runit it will be a big contender for sure.

It's just a word I know, but I don't think of a VoidPup as a 'contender' because that makes it sound like a competition, which I don't think is or should be how we think of forum distros - rather, all forum distros become alternative equal candidates for community versions/support - they are all welcome to me. All such community distro candidates end up using components from each other and that is good for each and all of us.

I too am really looking forward to a VoidPuppy public release, not because I want PPM (I don't) but because I'm keen to use Void's official package manager xbps with it - and then it should be very reliable using Void packages too, and Puppy is famous for its efficiency in terms of size (which is particularly great for running in RAM). Certainly, 'size' no longer matters in many scenarios, but sometimes it does (little Virtual Machines for example, can be run without using up too much in the way of resources).

[wiak]

Octoxbps works great.
Very simple GUI, but it does give visual listing of available stuff and indication if installed or not.
If you stick with xbps.
This should be how to run it.

However, now that I am not root.

Got to use sudo to even run wiakwifi to get network connection.

Got to use:

Code: Select all

sudo loadmodulegui

in terminal, to get SFS load GUI to run.

Overall, as weedog, a lot of programs seem easier to use.

Thurnar
Default applications seem to be recognized better, as the one to use for a specific file, you click on.

The drive icons on desktop
Seem to know what to do when clicked on.

I installed ntfs-3g package
So ntfs format can be accessed.

Nowadays there are advantages in running a distro as a normal user, for sure, used to be just an argument about security - which was never strong enough since we run as root user for years most everyone here and no security issues... But other issues to do with apps not being designed to run as root nowadays. However, we can still run as root, with workarounds same as always done. Alternatively can run as normal user, which I often now do: but, I do not use sudo for every command or I'd go mad!!!

I already tried to explain how to easily run as root when you need to. Basically I just start Thunar as root user with sudo, but then everything started from that Thunar instance will be as root user and work from that point of view - for example, open a terminal from Thunar right-click "Open Terminal Here" command and you'll find it opens as a root user terminal and apps will start with correct Xauthority form inside it (so no issues with X apps not starting up). It's easy, but sure, alternatively just stay as root desktop user most of the time and only switch to normal user (such as weedog or spot) when you need to use commands that require that (such as octoxbps); afterall how often will you actually end up needing to install a new package once you have things how you want it?

Code: Select all

sudo thunar

is the only sudo instance you need...

------------------------------------------------------------------------------------------------------
Said it already, though nothing to do with me. KLV is not a Puppy (but I think you already know this so what would be you reason for asking). It does not come from woof-CE build scripts. It is, I guess, named Kennel Linux because it is intended to be a distro using components from the various distros discussed on the forum. These individual distros such as Puppy and the DebianDogs continue to be developed in their own pure forms. WeeDogLinux itself isn't actually any particular distro (though it can be used to make one). It's more of a build system (a different one from woof-CE and not at all derived from it) and it includes an independent very generic initrd that is able to provide frugal install facilities (currently via overlay filesystem) for save persistence sfs module-handling and the like. It's not Puppy init though, which provides similar things.
KLV-airedale uses puppy huge kernel for its advantageous effect - that part is indeed a Puppy component created by woof-CE kernel kit I believe.
KLV's root filesystem was built using a WDL build script, but the result of that is fully Void Linux (to start with, since using xbps and upstream Void repos) in terms of compatibility, albeit specially chosen packages and special configurations by rockedge, and the rest of those providing input here (including already now extra functionality such as sfs-load system by fredx181 from the DebianDogs) and more such coming besides.
No-one could or would ever prevent anyone producing a Kennel Linux hybrid that uses a Puppy initrd, or a Puppy root filesystem (from woof-CE) or a DebianDog initrd and so on mixed with parts of other distros that were found to be useful. Daft not to use whatever is available on the forum etc, wherever, really.

I can't understand the question about what is KLV-airedale really, or see any importance in categorising it further. It is whatever it is or becomes via community feedback and input. Other KLV distros are inevitably going to be designed in different shapes and forms and component choices. It is Kennel Linux, a reflection or hybrid if you like of the various distros (including Puppy) in this Kennel.

Actually IMHO useful development time is always just wasted by discussions about "what is it exactly?". It is a distro to be developed and used or not at user-discretion - that's all. It involves the whole community in terms of "make it what you will".

A side-effect/benefit of KLV-airedale, for Puppy-enthusiasts, I suppose, is that a VoidPup is also now under development - that one is a Puppy (in that it is build via woof-CE apparently) though it is allowing experiments of using Void Linux official package manager (xbps) with Puppy too (instead of PPM). Development work on KLV-airedale will thus also benefit understanding of xbps and so on on any VoidPup release. Component parts of that could be used to build alternative Kennel Linux variants though - maybe via DebianDog parts or WeeDogLinux parts or whatever. Kennel Linux is Kennel Linux - I think that is obvious by its name.

Even if someone ends up using VoidPup in preference, their time helping to develop the large community-representing KLV-airedale will not be wasted. Much learned on KLV-airedale will be immediately applicable to any VoidPup and especially if it uses xbps. You may also learn about using runit as a service management methodology (a good alternative to systemd and to old sysVinit) and who knows that that will not also become a future improvement for Puppy?

I'll say no more about what I think Kennel Linux is though. Like I say, the definition/categorisation means nothing to me. Think of it as you will. I don't think rockedge needs to justify its creation at all.