Bookworm Build script

[trister]

Thanks for the responce fred.
This is the first dog that was almost 100% "drop-in" replacement of my bullseye system. I just changed 01-filesystem and the kernel files and the system booted. This is the easiest upgrade process ever

@trister

I used the appimage

Tested making a build with mklive-bookworm64 (appimage) using your attached config "build_setup__240730bookworm02.conf.txt" and it boots fine for me, so, no idea what can be the problem for you.

The system boots ok when using i915.modeset=1 .

It seems to me like it doesn't produce output at the start of the boot process (during the part which it loads initrd , squashfs files etc). It could send signal to HDMI instead of VGA or something like that and it then does the correct detections.
Is there a way to force detection or output to the screen from the start of the boot proccess .

I planned to do it after you added the option for non-merged filesystem.

Btw, I noticed in your config that it's containing NMU="TRUE" (no-merged-usr) but the appimage can only build with merged-usr.
The mklive-bookworm script from first post does have the choice for no-merged-usr though .

EDIT: You may want to try booting with a different (older?) kernel, but be sure it's usr-merged or no-usr-merged matching with the filesystem.

It seems to me that I got a non-merged filesystem.I used the first file from the topic and renamed .gz to appaimage.

[fredx181]

It seems to me like it doesn't produce output at the start of the boot process (during the part which it loads initrd , squashfs files etc). It could send signal to HDMI instead of VGA or something like that and it then does the correct detections.
Is there a way to force detection or output to the screen from the start of the boot proccess .

Understanding better now what you mean, it may have to do with systemd boot.
With systemd boot, I see the squashfs's loading for a very short while, when it's done, the console seems to get cleared and only the output of the systemd boot process shows for me. (very verbose and fast running)
(different as when booting without systemd, then the console isn't cleared after the squashfs loading).

I used the first file from the topic and renamed .gz to appaimage.

Ah ok. There was a bit confusion, the attachment at first post is a script, the appimage download link is here: https://debiandog.github.io/MakeLive/Re ... rmdog.html

[dancytron]

Working on and off on an LXQt version.

Use this with the default "skip" choice gets you to a working desktop with lots of missing icons and working sound.

Code: Select all

DESK_APPS_INSTALL="lxqt-core slim lxqt"
  
FIRMWARE="firmware-linux-free firmware-realtek firmware-linux-nonfree firmware-iwlwifi firmware-misc-nonfree vlc pipewire-setup"

My first attempt was to use "recommended=true", which included a ridiculous amount of unneeded stuff.

Just lxqt-core gets you to where you choose a desktop session but there aren't any.

[fredx181]

Working on and off on an LXQt version.

Use this with the default "skip" choice gets you to a working desktop with lots of missing icons and working sound.

Code: Select all

DESK_APPS_INSTALL="lxqt-core slim lxqt"
  
FIRMWARE="firmware-linux-free firmware-realtek firmware-linux-nonfree firmware-iwlwifi firmware-misc-nonfree vlc pipewire-setup"

My first attempt was to use "recommended=true", which included a ridiculous amount of unneeded stuff.

Just lxqt-core gets you to where you choose a desktop session but there aren't any.

Yes, tried exactly as you did and indeed many missing icons.
For Pcmanfm-qt and Desktop icons I could fix it by setting the icon theme from Menu > Preferences > Lxqt settings > Appearance .
But for the menu icons missing (mostly categories) I don't know.
The thing is that it would need a similar config setting addition as I've made earlier for some other Desktop Environments, don't know much about LXQT to be honest.

[dancytron]

Working on and off on an LXQt version.

Use this with the default "skip" choice gets you to a working desktop with lots of missing icons and working sound.

Code: Select all

DESK_APPS_INSTALL="lxqt-core slim lxqt"
  
FIRMWARE="firmware-linux-free firmware-realtek firmware-linux-nonfree firmware-iwlwifi firmware-misc-nonfree vlc pipewire-setup"

My first attempt was to use "recommended=true", which included a ridiculous amount of unneeded stuff.

Just lxqt-core gets you to where you choose a desktop session but there aren't any.

Yes, tried exactly as you did and indeed many missing icons.
For Pcmanfm-qt and Desktop icons I could fix it by setting the icon theme from Menu > Preferences > Lxqt settings > Appearance .
But for the menu icons missing (mostly categories) I don't know.
The thing is that it would need a similar config setting addition as I've made earlier for some other Desktop Environments, don't know much about LXQT to be honest.

I unfortunately have a thousand things going on right now, so not much time to spend on it.

I have the default debian bookworm lxqt version firstribed and working. I'll try to take a look and see if I can figure out something.

[fredx181]

But for the menu icons missing (mostly categories) I don't know.

Did spend a few hours trying to make all icons show in the menu (possible solutions from a web search I tried too), but no luck.
It seems to me that only /usr/share/pixmaps and /usr/share/icons/hicolor is being searched for icons, not the icon-theme that's configured as default theme.
Pity, on itself I like the LXQT DE, but the icons problem.... (I may be missing something (perhaps simple) though).

[trister]

It seems to me like it doesn't produce output at the start of the boot process (during the part which it loads initrd , squashfs files etc). It could send signal to HDMI instead of VGA or something like that and it then does the correct detections.
Is there a way to force detection or output to the screen from the start of the boot proccess .

Understanding better now what you mean, it may have to do with systemd boot.
With systemd boot, I see the squashfs's loading for a very short while, when it's done, the console seems to get cleared and only the output of the systemd boot process shows for me. (very verbose and fast running)
(different as when booting without systemd, then the console isn't cleared after the squashfs loading).

The issue seems to be motherboard/graphics card related. The same frugal install on my i3 shows everything during boot.

Also, I now have another issue.
I have an rtl 8821CU USB Wifi. In bullseye I had managed to compile it for kernel k-6.1.0-0.deb11.13-amd64 (I compiled from here : https://github.com/morrownr/8821cu-20210916 ). In boomworm I can't compile it because it says it expects a merged filesystem and bin folder should be a link...
In bookworm I use k-6.1.0-23-amd64-nmu. Is there something I can do -expept move everything to a merged filesystem? Can someone else compile it and send me an Squashfs?
Also, with compiling drivers I have the problem that I can't replace kernel without recompiling (and for backport kernels I could never get the correct dependencies).

Is there a newer kernel for bookworm I can use that already contains this driver?
Sorry for bothering you for these type of technical stuff but the driving compile is the biggest problem for me -since Its not a simple to make a squashfs of it.

[dcung]

Also, I now have another issue.
I have an rtl 8821CU USB Wifi. In bullseye I had managed to compile it for kernel k-6.1.0-0.deb11.13-amd64 (I compiled from here : https://github.com/morrownr/8821cu-20210916 ). In boomworm I can't compile it because it says it expects a merged filesystem and bin folder should be a link...
In bookworm I use k-6.1.0-23-amd64-nmu. Is there something I can do -expept move everything to a merged filesystem? Can someone else compile it and send me an Squashfs?

I have RTL8812BU which I have to recompile the driver, each time I refresh my DDOG iso sets - for different kernels.
So, I'm familiar with the routine. But instead of squashfs, I make .deb file.
https://github.com/morrownr/88x2bu-20210702

I downloaded and compiled yours in my fresh bookworm k-6.1.0-23-amd64-nmu
I didn't get errors. But I have no HW to test it.

No warranty, use at your own risk - Here is the .deb file if you'd like to test. Install the .deb - reboot.
https://drive.google.com/file/d/1ze_g_D ... sp=sharing

[trister]

Also, I now have another issue.
I have an rtl 8821CU USB Wifi. In bullseye I had managed to compile it for kernel k-6.1.0-0.deb11.13-amd64 (I compiled from here : https://github.com/morrownr/8821cu-20210916 ). In boomworm I can't compile it because it says it expects a merged filesystem and bin folder should be a link...
In bookworm I use k-6.1.0-23-amd64-nmu. Is there something I can do -expept move everything to a merged filesystem? Can someone else compile it and send me an Squashfs?

I have RTL8812BU which I have to recompile the driver, each time I refresh my DDOG iso sets - for different kernels.
So, I'm familiar with the routine. But instead of squashfs, I make .deb file.
https://github.com/morrownr/88x2bu-20210702

I downloaded and compiled yours in my fresh bookworm k-6.1.0-23-amd64-nmu
I didn't get errors. But I have no HW to test it.

No warranty, use at your own risk - Here is the .deb file if you'd like to test. Install the .deb - reboot.
https://drive.google.com/file/d/1ze_g_D ... sp=sharing

Thanks , I tried it but it didn't work (I use frisbee btw). Your deb contains very few files and does no system file modifications (this is what I was trying to achieve ). It would be ideal if this worked - even if I had to add a bash script to activate it (but I don't have the knowledge).

I made a non-merged system and compiled the driver . I made a squashfs with as few files possible (the /opt folder is just a clone of the github repo) . https://app.prismdrive.com/drive/s/ByXp ... kr6zVF9WRH .
I believe the work is done by the /usr/lib/modules/6.1.0-23-amd64/modules* files (which overwrite my 01-filesystem files -which I would like to avoid). Your deb did not contain that so this might be why it didn't work for me (maybe in your system these files were modified during the compile/install process).

Just a note I use the EXIT: parameter so I never keep changes in my save file. I make squashfs with the modification I need and add this to my system.

[dcung]

Thanks , I tried it but it didn't work (I use frisbee btw). Your deb contains very few files and does no system file modifications (this is what I was trying to achieve ). It would be ideal if this worked - even if I had to add a bash script to activate it (but I don't have the knowledge).

I made a non-merged system and compiled the driver . I made a squashfs with as few files possible (the /opt folder is just a clone of the github repo) . https://app.prismdrive.com/drive/s/ByXp ... kr6zVF9WRH .
I believe the work is done by the /usr/lib/modules/6.1.0-23-amd64/modules* files (which overwrite my 01-filesystem files -which I would like to avoid). Your deb did not contain that so this might be why it didn't work for me (maybe in your system these files were modified during the compile/install process).

Just a note I use the EXIT: parameter so I never keep changes in my save file. I make squashfs with the modification I need and add this to my system.

The actual (compiled) files needed are (these 2 files in the .deb file).
/etc/modprobe.d/8821cu.conf
/lib/modules/6.1.0-23-amd64/kernel/drivers/net/wireless/realtek/8821cu.ko

In my case, for my USB WiFi, I can just copy them manually and reboot, and it works.
PeasyWiFi or frisbee..both work in my case once the driver is loaded.

But if you don't keep the changes, I don't think it will work, since reboot is needed to activate the driver.

Maybe if you make squashfs with those two files (?).

[fredx181]

@trister
According to what's discussed here (see last post specially) you can use the backports kernel and install realtek firmware deb from Trixie
https://plethora.debian.net/viewtopic.p ... a7#p800483

Here's backports kernel 'package' (nmu) that I made using 'upgrade-kernel' (first added the bookworm backports repo in sources.list)
https://github.com/doglinux/bookworm/re ... nmu.tar.gz
firmware-realtek deb (from Trixie):
https://packages.debian.org/trixie/all/ ... k/download

[trister]

But if you don't keep the changes, I don't think it will work, since reboot is needed to activate the driver.

yes, that seems to be the problem. During compile there are some extra changes happening to the filesystem.

Maybe if you make squashfs with those two files (?).

The squashfs I made by checking the changes made in the filesystem when I compiled and installed the driver. I believe the modules.bin is an important part.

@trister
According to what's discussed here (see last post specially) you can use the backports kernel and install realtek firmware deb from Trixie
https://plethora.debian.net/viewtopic.p ... a7#p800483

Here's backports kernel 'package' (nmu) that I made using 'upgrade-kernel' (first added the bookworm backports repo in sources.list)
https://github.com/doglinux/bookworm/re ... nmu.tar.gz
firmware-realtek deb (from Trixie):
https://packages.debian.org/trixie/all/ ... k/download

Thanks fred, I will try those

Btw, I created a merged system. I needed to change only 5-6 of my files and it seems to work. (Assuming that the only difference are the bin,lib,lib64,sbin that are linked inside /usr.
(Thats how I managed to compile the driver cause of the merged system),

[fredx181]

Btw, I created a merged system. I needed to change only 5-6 of my files and it seems to work. (Assuming that the only difference are the bin,lib,lib64,sbin that are linked inside /usr.

If you did that manually (by editing filesystem,squashfs to make it merged) I'm not sure if that's good, may give complications (not sure though )

[trister]

Btw, Thanks Thanks fred.
I created a backport kernel and included the trixie firmware and it worked (without compiling or making sfs etc).

Btw, I created a merged system. I needed to change only 5-6 of my files and it seems to work. (Assuming that the only difference are the bin,lib,lib64,sbin that are linked inside /usr.

If you did that manually (by editing filesystem,squashfs to make it merged) I'm not sure if that's good, may give complications (not sure though )

I never edit 01-filesystem. I just make changes and load them after 01-fileystem so it's data is overwritten. I know the implications of that . in my early days I used apt2sfs made files . Now if I have to use an apt2sfs file I strip the parts that I know cause problems.

[fredx181]

@trister Ah ! Great that it works for you with the backports kernel.

I never edit 01-filesystem. I just make changes and load them after 01-fileystem so it's data is overwritten. I know the implications of that .

I just thought maybe you manually did convert a no-merged-usr system into a merged-usr system (which would be tricky IMO).
Nothing wrong basically with editing 01-filesystem, I'd say, it just depends, I use no extra .squashfs files (besides the kernel .squashfs) and do once in a while a quick-remaster which adds my changes to new created 01-filesystem.squashfs (but I know, your usage is different) .

[trister]

(edit 2): I noticed that the main problem was caused my 2 squashsfs files with MANY files (21k and 15k - thet contained conty.sh & Wine prefixes with installed apps). This would be the answer to the problem but I was wondering why in Bulleye I had no problem with these ( I use these files for more than a year).
Is there something going with bookworm just before the start of the windows environment ? Does it read the file list of the sfs files?

-----The rest is more for details about my delay problem----

I noticed another issue during boot in my i3 PC.
During the boot process when the gui starts (after SLIM login manager) there is a 1minute and 10 seconds delay. During that delay screen goes black and mouse cursor and GUI appears after that delay. Delay is the same when I have Wifi USB connected and when I remove it (and have no internet)

The delay I think is between these entries in syslog:

Code: Select all

2024-08-13T12:53:01.952724+03:00 live start-tzupdate[1620]: No network connection yet...
...
2024-08-13T12:54:01.509332+03:00 live dbus-daemon[2337]: [session uid=0 pid=2335] Successfully activated service 'org.xfce.Xfconf'

You can view my full syslog here : https://pastebin.com/CKUhm1ZN

I think it might have something to do with tzupdate which doesn't allow the graphical environment to start. (I'll try to post a video but my time is limited)

Any idea what might cause the delay?

edit :
During that black screen delay the disk seems to constantly read stuff.
-It delays 1m10seconds from my external USB 3 SSD
-When I run in from my internal M2 it delays 30 seconds
-In a clean installation (with 5 squashfs it takes 10-15seconds )

Is it possible that the system reads the squashfs contents of all the files during that time ( I thought it read them during the time it displayed their name earlier)?

[fredx181]

Is there something going with bookworm just before the start of the windows environment ? Does it read the file list of the sfs files?

The sfs files are mounted and added to the overlay stack at the init stage, very unlikely IMO that it's 'read' at a later stage (and nothing I changed for that compared to Bullseye).
I think I have a very similar build as you (made from your .conf that you shared above as attachment,( i.e. DDog DE with systemd, no-merged-usr) and booting it, I cannot reproduce the long delay, for me, after the slim login, the Desktop appears almost instantly. (EDIT: I have no ethernet too on my laptop, relying on WIFI only)
So I don't know, sorry. I guess that the only way for you to find the cause is to experiment, e.g. remove package tz-update, boot without extra modules etc...
Did you have the delay problem also before using the backports kernel + trixie firmware ?

edit: I'm no good at getting conclusions from the log you attached , it seems to me a combination of things that cause the delay, dhcpcd, systemd
Perhaps it's time to replace frisbee (has old dhcpcd version included) with a more up-to-date network-manager (network-manager-gnome perhaps. edit: or peasywifi may work better in this case) .

[trister]

The sfs files are mounted and added to the overlay stack at the init stage, very unlikely IMO that it's 'read' at a later stage (and nothing I changed for that compared to Bullseye).

Is it possible that my system is too quick and we reach a point where it tries to show the desktop but it still tries to read the contents of the squashsfs files? (maybe my external USB 3 ssd is slow)?
Why I say that? Because booting the exact same frugal installed folder from my internal M.2 disk, the (black screen) delay is always exactly half . (Eg from 20seconds is 10 seconds. From 60secinds with all files it was 30seconds).

I think I have a very similar build as you (made from your .conf that you shared above as attachment,( i.e. DDog DE with systemd, no-merged-usr) and booting it, I cannot reproduce the long delay, for me, after the slim login, the Desktop appears almost instantly. (EDIT: I have no ethernet too on my laptop, relying on WIFI only)

In an almost clean frugal install the delay is about 10 seconds or less - I wouldn't consider this a delay.
I have pinpointed the (biggest) cause of the delay to the 2 big SFS ( 1.5GB+ size with 21k and 15k files each. I added only one of these with half the files and I am now in a 20 seconds delay.

Did you have the delay problem also before using the backports kernel + trixie firmware ?

I must setup a frugal install at some point to check that but I think I had this in all bookworm kernels I tried.

Perhaps it's time to replace frisbee (has old dhcpcd version included) with a more up-to-date network-manager (network-manager-gnome perhaps) .

network-manager-gnome loads extra gnome libraries?
Can you point me to any other better network managers to check?

[fredx181]

The sfs files are mounted and added to the overlay stack at the init stage, very unlikely IMO that it's 'read' at a later stage (and nothing I changed for that compared to Bullseye).

Is it possible that my system is too quick and we reach a point where it tries to show the desktop but it still tries to read the contents of the squashsfs files? (maybe my external USB 3 ssd is slow)?
Why I say that? Because booting the exact same frugal installed folder from my internal M.2 disk, the (black screen) delay is always exactly half . (Eg from 20seconds is 10 seconds. From 60secinds with all files it was 30seconds).

I think I have a very similar build as you (made from your .conf that you shared above as attachment,( i.e. DDog DE with systemd, no-merged-usr) and booting it, I cannot reproduce the long delay, for me, after the slim login, the Desktop appears almost instantly. (EDIT: I have no ethernet too on my laptop, relying on WIFI only)

In an almost clean frugal install the delay is about 10 seconds or less - I wouldn't consider this a delay.
I have pinpointed the (biggest) cause of the delay to the 2 big SFS ( 1.5GB+ size with 21k and 15k files each. I added only one of these with half the files and I am now in a 20 seconds delay.

Did you have the delay problem also before using the backports kernel + trixie firmware ?

I must setup a frugal install at some point to check that but I think I had this in all bookworm kernels I tried.

Perhaps it's time to replace frisbee (has old dhcpcd version included) with a more up-to-date network-manager (network-manager-gnome perhaps) .

network-manager-gnome loads extra gnome libraries?
Can you point me to any other better network managers to check?

I really have no useful comment about your issue, as I said, the extra .squashfs files are mounted and merged into the root filesystem / at init stage, could be that a very large filesystem causes the delay, not sure, but you didn't have that problem on Bullseye as far as I understand so... mystery to me

network-manager-gnome loads extra gnome libraries?
Can you point me to any other better network managers to check?

network-manager-gnome takes more space as frisbee, but it doesn't load extra gnome (desktop) stuff AFAIK, possibly worth to try peasywifi also, btw (edit: or connman).

[dancytron]

Working on and off on an LXQt version.

Use this with the default "skip" choice gets you to a working desktop with lots of missing icons and working sound.

Code: Select all

DESK_APPS_INSTALL="lxqt-core slim lxqt"
  
FIRMWARE="firmware-linux-free firmware-realtek firmware-linux-nonfree firmware-iwlwifi firmware-misc-nonfree vlc pipewire-setup"

My first attempt was to use "recommended=true", which included a ridiculous amount of unneeded stuff.

Just lxqt-core gets you to where you choose a desktop session but there aren't any.

Yes, tried exactly as you did and indeed many missing icons.
For Pcmanfm-qt and Desktop icons I could fix it by setting the icon theme from Menu > Preferences > Lxqt settings > Appearance .
But for the menu icons missing (mostly categories) I don't know.
The thing is that it would need a similar config setting addition as I've made earlier for some other Desktop Environments, don't know much about LXQT to be honest.

I used the brute force approach of installing every package that could possible have icons in them.

Code: Select all

DESK_APPS_INSTALL="lxqt-core slim lxqt lxqt-themes lxqt-system-theme papirus-icon-theme breeze-icon-theme faenza-icon-theme oxygen-icon-theme"

And it worked!!! Everything seems to have an icon.

Things seem generally to work and it's very fast. Sound works.

Dan

edit: and the basic tiling/snapping windows works too (which I've gotten very used to using MX on my laptop).

edit again: right click shortcut for filemnt and edit-sfs work.

For reference: https://wiki.debian.org/LXQt#Package_Manager

[dcung]

I used the brute force approach of installing every package that could possible have icons in them.

Code: Select all

DESK_APPS_INSTALL="lxqt-core lxqt lxqt-themes lxqt-system-theme papirus-icon-theme breeze-icon-theme faenza-icon-theme oxygen-icon-theme"

Curious what it looks like.... I currently tinkering on my fujitsu laptop, so using it to test it.
I build one using openbox.conf replacing DESK_APPS_INSTALL with what you had...(except slim).

Right-click Install .deb work - I used it to install deadbeef to test sound, works.
Default is 'double-click' (?) for things... probably configurable as with most DE...
Quick-remaster run - I didn't make one though...

2024-08-15-120506_1366x768_scrot.png (334.44 KiB) Viewed 1919 times

[dancytron]

Conky was broken on mine.

I started with the .conf file that was generated when you choose "skip", which I think is why conky was broken on mine.

I used it for a couple of hours last night for general video and music and web surfing on my mid- Windows 7 era desktop with 4 gig of ram and it was fast and stable.

[fredx181]

Working on and off on an LXQt version.

Use this with the default "skip" choice gets you to a working desktop with lots of missing icons and working sound.

Code: Select all

DESK_APPS_INSTALL="lxqt-core slim lxqt"
  
FIRMWARE="firmware-linux-free firmware-realtek firmware-linux-nonfree firmware-iwlwifi firmware-misc-nonfree vlc pipewire-setup"

My first attempt was to use "recommended=true", which included a ridiculous amount of unneeded stuff.

Just lxqt-core gets you to where you choose a desktop session but there aren't any.

Yes, tried exactly as you did and indeed many missing icons.
For Pcmanfm-qt and Desktop icons I could fix it by setting the icon theme from Menu > Preferences > Lxqt settings > Appearance .
But for the menu icons missing (mostly categories) I don't know.
The thing is that it would need a similar config setting addition as I've made earlier for some other Desktop Environments, don't know much about LXQT to be honest.

I used the brute force approach of installing every package that could possible have icons in them.

Code: Select all

DESK_APPS_INSTALL="lxqt-core slim lxqt lxqt-themes lxqt-system-theme papirus-icon-theme breeze-icon-theme faenza-icon-theme oxygen-icon-theme"

And it worked!!! Everything seems to have an icon.

Things seem generally to work and it's very fast. Sound works.

Dan

edit: and the basic tiling/snapping windows works too (which I've gotten very used to using MX on my laptop).

edit again: right click shortcut for filemnt and edit-sfs work.

For reference: https://wiki.debian.org/LXQt#Package_Manager

Thanks Dan! Icon problem solved.
Investigated some more and I've found that only installing papirus-icon-theme will do (other icon themes not needed to show icons in the Menu).
Strange thing is still that when selecting another icon-theme from settings, the menu icons won't change (stays papirus icons), well ... no real problem though.

[fredx181]

Experimented adding "Lxqt basic" to the DE Choices (only with -gui for now), I think LXQT is a keeper, mklive-bookworm-exp script :

mklive-bookworm-exp.gz

Remove fake .gz and make executable

(72.12 KiB) Downloaded 35 times

2024-08-15_17-20-33.png (153.62 KiB) Viewed 1849 times

Added edit-sfs-pcmanfm-qt and filemnt-pcmanfm-qt to the repo, included in the lxqt.conf (to prevent that pcmanfm is installed also (besides pcmanfm-qt) ), made small change to the conky config and added "Logout" on the Desktop (running "obshutdown", for in case choice required (graphical) to save changes or not).
Btw, I think there's a problem with slim, when installed, the choice to Save/NoSave from console doesn't work for me (I use the EXIT: option) (will save though, but cannot choose NoSave).

ISO just made from the LXQT build : https://github.com/doglinux/book-worm/r ... 4-UEFI.iso
The 'lxqt' package (included) adds more than maybe needed, e.g pulseaudio, pavucontrol, featherpad .....
Without it, it's possible too perhaps (with just lxqt-core installed), but needs xfwm4 to get to a proper Desktop (but even then it's a bit crippled, better have the "full" IMO).

Screenshot_840x472.png (482.08 KiB) Viewed 1834 times

[dancytron]

I saw the same thing at with the save at the end.

It felt like the keyboard was locked.

I'll try building one with a different session manager.

LightDM maybe.

I agree we want the full version.

edit: built one with lightdm instead of slim. It booted on my chromebook and save2flash seemed to work correctly when restarting. No sound on the chromebook otb, but that's to be expected.

Code: Select all

DESK_APPS_INSTALL="lxqt lightdm lxqt-core lxqt-themes lxqt-system-theme papirus-icon-theme tzupdate edit-sfs-pcmanfm-qt filemnt-pcmanfm-qt"
  
FIRMWARE="firmware-linux-free firmware-realtek firmware-misc-nonfree firmware-linux-free firmware-realtek firmware-linux-nonfree firmware-iwlwifi firmware-misc-nonfree vlc pipewire-setup

[fredx181]

I saw the same thing at with the save at the end.

It felt like the keyboard was locked.

Made small change in the snapexit script , console Save/NoSave should work ok now with slim installed, (at least it does on my laptop) . Upgrade porteusbootscripts package to version 1.1.9 (includes the new snapexit script).
edit: nice that it works well with lightdm (must say though that I don't like it much, not very "root friendly" from what I experienced).

[dancytron]

Building another one with qbittorrent, samba, yassm and the other stuff I'll want on my desktop/server.

A little weirdness with the pulse audio volume control. Probably an extra package somewhere.

edit: it seems like pipewire-setup includes pavucontrol when lxqt wants pavucontrol-qt.

All the pulse audio stuff seems to be installed and you can't uninstall it without getting rid of lxqt, but pipewire-setup is working as it usually does going straight from alsa to pipewire/wireplumber.

I guess the next thing to try is without pipewire-setup and see what it does by default.

[dcung]

I used this git-sync to pick @fredx181's lxqt.conf
viewtopic.php?p=126555#p126555

Modified a little to put some packages that I'm already used to.
I don't use session managers, no slim or lightdm for me.
I add alsamixergui and modified the panel volume to use it. Test that youtube has sound and can play few local mp3 files.
I like the battery indicator with percentage and set it to pop-up when low critical reached..
The C N S on the panel got me - accidentally clicked N, and NumLock gets activated unintentionally... I didn't know what they were for... I do now...

Even with obshutdown or EXIT option (saved)...
Some changes are still not saved (?)... eg. some computer.desktop, seems to revert back to what it was after reboot.
conkytoggle in Startup doesn't get executed since the 'Trust' reverted back (?) - so I had to manually click each reboot.
EDIT:
Turned out LXQT 'needs' gvfs. I installed gvfs and 'trust' issue went away.
https://github.com/lxqt/pcmanfm-qt/issues/737
https://github.com/lxqt/lxqt/issues/1687

But, I definitely will add 1 more DE/iso in the hoarding collection.
It's a painless few lines mod to my marathon builds.

2024-08-16-110535_413x463_scrot.png (52.87 KiB) Viewed 1760 times

[fredx181]

conkytoggle in Startup doesn't get executed since the 'Trust' reverted back (?) - so I had to manually click each reboot.

Strange... For me checking "Trust this executable" (right-click on desktop icon) does nothing more than removing the exclamation mark (but indeed the exclamation mark is back after reboot, so that setting is not saved ).
It cannot influence scripts in ~/Startup (in this case "conky_start"). For me conky starts everytime after reboot or logout (with some delay though, be patient ).

Btw, the best way to add icons to the Desktop is to go in "Applications" (from left side pane) (menu://applications/) and drag to the Desktop and choose "symlink". No exclamation mark then.

[dcung]

Strange... For me checking "Trust this executable" (right-click on desktop icon) does nothing more than removing the exclamation mark (but indeed the exclamation mark is back after reboot, so that setting is not saved ).
It cannot influence scripts in ~/Startup (in this case "conky_start"). For me conky starts everytime after reboot or logout (with some delay though, be patient ).

Btw, the best way to add icons to the Desktop is to go in "Applications" (from left side pane) (menu://applications/) and drag to the Desktop and choose "symlink". No exclamation mark then.

I should have mentioned/elaborated in my edit above.

Most of my builds, I put conkytoggle symlink in Startup to start it if not already there/start.

So I did the same in Lxqt. But the trust (i.e. exclamation mark) always come back after reboot. After testing a bit, I found that more than one way I can resolve it.
1. Create a bash script file (ie. #!/bin/bash type) - to execute conkytoggle. Sort of the conky_start you mentioned.
2. Install gvfs and the symlink no longer has exclamation mark. This is not just for conky, but anything I put on desktop, even after reboot.

I include gvfs in my build conf now, since I like to have it, I have a need for it too..so that issue is settled/resolved.