- Arch-general - lists.archlinux.org

btrfs send/receive between same UUID filesystems
by Alipour Alipour 25 Oct '25

25 Oct '25

Does having the same UUID for btrfs filesystem on local and remote machines cause problems when using btrfs send/receive between them (piped through SSH)?

2 1

clipit vs parcellite - Was: "Fwd: Remove GTK2 from the official repositories"
by Ralf Mardorf 24 Oct '25

24 Oct '25

On Fri, 2025-10-24 at 12:32 +0200, Robin Candau wrote: > If anything, gtk2 and related packages can still be maintained in the > AUR, but the point is we are not willing to put efforts into > officially maintaining such deprecated packages anymore. > > Please bear with us and thank you for your understanding :) It is understandable. Btw. I'm testing if clipit can replace parcellite. • root@archlinux /usr/bin # mv -i parcellite parcellite.ori && ln -s clipit parcellite Temporarily, just to see how far one could go. • rocketmouse@archlinux ~ $ grep parc .config/openbox/autostart (sleep 1 && parcellite -n) & Not really and this doesn't change when it's "correctly" startet via clipit -n. There are seemingly no options for the window position. Some windows appear at the cursor position, the menu appears at the upper right corner, but not after pushing <Ctrl><Alt>P, it appears after first pushing <Ctrl><Alt>P followed by pushing <Ctrl><Alt>F. • rocketmouse@archlinux ~ $ cat ~/.config/parcellite/parcelliterc [rc] RCVersion=1 use_copy=true use_primary=false synchronize=false save_history=true history_pos=true history_x=1000 history_y=800 history_limit=25 data_size=0 item_size=5 automatic_paste=false auto_key=false auto_mouse=true key_input=false restore_empty=true rc_edit=false type_search=false case_search=false ignore_whiteonly=false trim_wspace_begend=false trim_newline=false hyperlinks_only=false confirm_clear=true disable_clear=false save_history_Lineno=true current_on_top=true single_line=true reverse_history=false item_length=50 persistent_history=false persistent_separate=false persistent_on_top=false persistent_delim=\\n nonprint_disp=false ellipsize=2 multi_user=true on_unity=false icon_name=parcellite menu_key=<Ctrl><Alt>P history_key=<Ctrl><Alt>H phistory_key=<Ctrl><Alt>X actions_key=<Ctrl><Alt>A • rocketmouse@archlinux ~ $ cat ~/.config/clipit/clipitrc [rc] use_copy=true use_primary=false synchronize=false automatic_paste=false show_indexes=false save_uris=true use_rmb_menu=false save_history=false history_limit=25 history_timeout_seconds=30 history_timeout=false items_menu=20 statics_show=true statics_items=10 hyperlinks_only=false confirm_clear=false single_line=true reverse_history=false item_length=50 ellipsize=2 history_key=<Ctrl><Alt>H actions_key=<Ctrl><Alt>A menu_key=<Ctrl><Alt>P search_key=<Ctrl><Alt>F offline_key=<Ctrl><Alt>O offline_mode=false exclude_windows= Let's see if I can get used to it. Hm? Maybe copying quotes from the Evolution MUA works better with clipit.

1 1

[arch-dev-public] Re: Remove GTK2 from the official repositories
by mpan 24 Oct '25

24 Oct '25

> I think we've reached the point when we can remove the very old GTK2 > package from the official repositories. I recently dropped the GTK2 > version of LXDE, and there are only a few applications left that still > depend on GTK2: > (… pkglist …) > Most of these have no active development a long time ago. If no one has > objections, I'll drop these packages to AUR in the next days if it's > not possible to build them without GTK2. Hello, thanks for building the list. NOTE: steam-native-runtime is missing from the list! Shifting away from GTK2 is a good idea. And any package package that doesn’t build due to gtk2 dependency and is not in active development, should IMO be removed. But other than that I also believe due care is needed and there is no haste. Unless of course gtk2 itself starts to be a serious burden or is at risk of becoming orphaned. Many of the packages listed are not niche. To name a few that instantly caught my eye: pidgin, leafpad, hexchat, gimp-nufraw. To limit subjective perspective’s effect on my opinion, I checked with pkgstats.[tab:1] So this isn’t just a distortion from my singular vantage point. While some projects are either dead (hexchat) or dormant/stalled (leafpad, gimp-nufraw), others are active (pidgin). Being in the first group is not making a package obsolete either, even if users should think about searching for alternatives. If all these packages are moved to AUR, what happens with the now unneeded gtk2 package? Cheers, mpan [tab:1] pkgstats results. [WARNING: statistical data ahead! Values have to be interpreted with proper understanding of their meaning, context, and influences!] Package Popularity [%] -------------------------------- steam-native-runti 12.84 calf 12.04 hexchat 5.30 pidgin 5.30 nitrogen 5.01 bless 3.79 leafpad 3.51 gimp-nufraw 2.65 mono-tools 2.12 gtk-chtheme 2.10 amsynth 2.05 xboard 1.82 eq10q 1.79 gtk2fontsel 1.73 gkrellm 1.62 parcellite 1.57 hexter 1.28 xdialog 1.18 trayer 1.00 xlockmore 0.92 osdlyrics 0.71 alltray 0.55 nspluginwrapper 0.50 gpsim 0.46 sbxkb 0.30

1 0

Any particular reason the gtk2 build was dropped from libfm-gtk2 (int included both gtk2 and gtk3)
by David C Rankin 22 Oct '25

22 Oct '25

All, The latest update to libfm-gtk2 renames it libfm-gtk3 but make is conflict with existing libfm-gtk2. https://gitlab.archlinux.org/archlinux/packaging/packages/libfm/-/commits/m… The libfm-gtk2 package included both gtk2 and gtk3 file manager builds. Now we are left with only gtk3. It was quite convenient for old apps that use the gtk2 functionality. Why did the gtk2 build get dropped? It was working quite nicely with the gtk3 filemanager files, posed no conflict and no additional work. Yes, it's old, but so are a lot of things that work quite well. If there weren't build failures, what was the reason behind the change? -- David C. Rankin, J.D.,P.E.

2 2

Btrfs and nspawn
by Pocket 28 Sep '25

28 Sep '25

I have a system that uses btrfs for the root filesystem. I created a subvolume root which becomes the root filesystem on boot up, ie it bootds to the root subvolume. I now want to add some "containers" to that machine, for example DNS. When set up with bind and started with machinectl provides DNS for the network. Is it possible to use a snapshot for the root subvolume for nspawn containers? I have seen that one can use the --template option but it appears to be a newly created container every time. Thsat won't work as I would lose the bind setup. I need something that is not volatile in which I can install bind an configure it without having to do that every time the container is shutdown and the restarted. Also if I can use the root subvolume as a template will the subsequent containers be updated when I use pacman -Syu to update the root subvolume? -- Hindi madali ang maging ako

2 2

Btrfs and nspawn
by Pocket 28 Sep '25

28 Sep '25

-------- Forwarded Message -------- Subject: Re: Btrfs and nspawn Date: Sun, 28 Sep 2025 05:48:04 -0400 From: Pocket <pocket(a)columbus.rr.com> To: Moabeat <moabeat(a)berlin.de> On 9/28/25 2:45 AM, Moabeat wrote: > Let me check if I understood this correctly: > > 1. You want to set up a container for prividing DNS services. > 2. You want this setup of DNS to persist between container restarts. > 3. You want the container to keep up with the system updates of your > host system. Yes that is correct. > > Maybe it dind't get the point of what you try to achieve, but for me > this raises the question why you want to use a container at all? IMHO > the best option would be, to just install the packages for the DNS > services right away in your host, configure them and ditch the > container. The whole point of a container is, that it can be run and > updated independently of the host. Or is there anything I missed? I am doing this already on a host that uses ext4 for the filesystem and I wish to change that to btrfs. By using a container (systemd-nspawn) it allows me to keep the service separate from the host system. I then can copy that to move the container to another host without trouble. I am doing this on a single host that has containers for DNS, Nginx and dovecot/opensmtpd. I want to snapshot the "root" subvolume as the starting point for the container. If I understand the process the I won't have to "install" the "rootsystem" in each container and they will share those files. The only differences will be whatever changes need to be made for each container saving a bunch of SDD space. When setting up the "subvolumes/containers" for each service if I use the --template option it appears to do that but it is voltile as it will setup a new subvolume each time that will a "copy" of the the root subvolume, without bind and settings. I originally though I could do this by snapshotting the "root" subvolume and then use that for the systemd-nspawn container starting it with machinectl upon system boot. I have not tried that. I am asking if anyone here has done something like this. Looking for pointers information. > > On Sun Sep 28, 2025 at 2:24 AM CEST, Pocket wrote: > >> I have a system that uses btrfs for the root filesystem. >> >> I created a subvolume root which becomes the root filesystem on boot >> up, ie it bootds to the root subvolume. >> >> I now want to add some "containers" to that machine, for example DNS. >> >> When set up with bind and started with machinectl provides DNS for the >> network. >> >> Is it possible to use a snapshot for the root subvolume for nspawn >> containers? >> >> I have seen that one can use the --template option but it appears to >> be a newly created container every time. >> >> Thsat won't work as I would lose the bind setup. >> >> I need something that is not volatile in which I can install bind an >> configure it without having to do that every time the container is >> shutdown and the restarted. >> >> Also if I can use the root subvolume as a template will the subsequent >> containers be updated when I use pacman -Syu to update the root >> subvolume? > -- Hindi madali ang maging ako

4 4

/boot full
by Brian Keck 26 Sep '25

26 Sep '25

Last time (24 Aug) I updated arch on my laptop, /boot filled up. [1] There was a message saying that the build of /boot/initramfs-linux-fallback.img failed [1] [2] seems to say to adjust /etc/mkinitcpio.conf by dropping kms from HOOKS & adding i915 to MODULES. But this makes me nervous, so I'm asking here. Laptop has only Intel graphics [3], & lsmod shows both i915 & xe. I'm using both X11 & Wayland. Is the above advice good? Should I add xe to MODULES as well? TIA, Brian Keck ------------------------------------------------------------------------ [1] Laptop is Lenovo ThinkPad X1 Carbon Gen 12 14" /boot partition is 256 MB /boot/initramfs-linux-fallback.img size is 179 MB [2] https://bbs.archlinux.org/viewtopic.php?id=293333 [3] lspci -v 00:02.0 VGA compatible controller: Intel Corporation Meteor Lake-P [Intel Arc Graphics] (rev 08) (prog-if 00 [VGA controller]) Subsystem:Lenovo Device 231e Flags:bus master, fast devsel, latency 0, IRQ 166, IOMMU group 0 Memoryat 4058000000 (64-bit, prefetchable) [size=16M] Memoryat 4000000000 (64-bit, prefetchable) [size=256M] ExpansionROM at 000c0000 [virtual] [disabled] [size=128K] Capabilities:<access denied> Kerneldriver in use: i915 Kernelmodules: i915, xe

5 6

Sway Crashes After Recent Updates
by Radosław Kintzi 24 Sep '25

24 Sep '25

Not sure if this is the righ forum to report this kind of problems, but after this morning's `pacman -Syu`, I began to observe that Sway crashes each time DPMS is forced off, for example after `swaymsg "output * dpms off"`. Downgrading wlroots0.19 (0.19.1-2 => 0.19.0-5) and libdisplay-info (0.3.0-1 => 0.2.0-2) (and mpv) "fixes" the problem. Small excerpt from the system journal: wrz 24 10:07:01 ganges sway[1030]: sway: backend/drm/drm.c:1453: realloc_crtcs: Assertion `conn->crtc != NULL' failed. wrz 24 10:07:01 ganges systemd-coredump[1262]: Process 1030 (sway) of user 1000 terminated abnormally with signal 6/ABRT, proc> wrz 24 10:07:01 ganges systemd[1]: Created slice Slice /system/systemd-coredump. wrz 24 10:07:01 ganges systemd[1]: Started Process Core Dump (PID 1262/UID 0). wrz 24 10:07:01 ganges systemd-coredump[1263]: [🡕] Process 1030 (sway) of user 1000 dumped core. Best, R

3 4

failing to preform offline installation
by daggs 20 Sep '25

20 Sep '25

Greetings, I have a server that serves as router via vm, I'm attempting to install arch offline. here is what I did, I've created a json config to load using archinstall to preform the actual install. after selecting all the pkgs needed, I've extracted them from the json file and ran this: pacman -Syw --cachedir /mnt/repos/cache --dbpath /mnt/repos/db ..... now I have all the pkgs I need, next I've entered /mnt/repos/cache and ran this: repo-add ./custom.db.tar.zst ./*[^sig] according to the docs, I have local repo if I understand correctly. next I've booted the system in question and updated /etc/pacman.conf and added this: [custom] SigLevel = PackageRequired Server = file:///mnt/repos/cache I ran archinstall, got this: Fetching Arch Linux package database... Failed to sync Arch Linux package database. Most likely due to a missing network connection or DNS issue. Run archinstall --debug and check /var/log/archinstall/install.log for details. how can I instruct archinstall to use my local repo? Thanks.

1 0

How to stop email notifications from gitlab.archlinux.org
by Uģis Ģērmanis 19 Sep '25

19 Sep '25

Hello, There was a time I wanted all gitlab activity sent to my email, so i set Preferences => Notifications => Global notification level => Custom (to the activities I was interested) Now times have changed and I no longer have to the time to catch up with thousands of emails a week. So i tried setting the activity level to "Participate" which should only notify about things I'm involved in somehow, like issues I've commented on or things I've explicitly set to be notified about. But I still receive emails about issues created mostly by Bumpbuddy (and some others). Like hundred emails a day. Also ticking and unticking "Receive notifications about your own activity' doesn't seem to change anything either. I also tried to set the global notification level to "Disabled" which should kill all emails as it says but even that has no effect either. If I do a search on issues " Subscribed = Explicitly subscribed", I only get 1 closed issue. So I believe something is broken, because I haven't explicitly subscribed to anything that much, so the global setting shouldn't be bypassed. ugjka

1 1