- Arch-general - lists.archlinux.org

mkarchroot fails with "Permission denied" error during pacman sync (maybe alpm download user related?)
by Henning 25 Sep '24

25 Sep '24

Hi, I hope this is (one) of the correct places to request some help. I followd this Wiki article (https://wiki.archlinux.org/title/DeveloperWiki:Building_in_a_clean_chroot) and failed on the classical way and I have no clue why :-) I add some example output on the buttom. Maybe related. The "arch machine" with this error is a VM, running via incus on an arch host. The VM use ext4, the host btrfs. On the arch host itself, the command work as expected. Only inside the VM it fails. If I comment out "DownloadUser = alpm" in /etc/pacman.conf, it works. But this is also enabled on the (working) host, without any problems. (same default pacman.conf) Thanks Henning ################################################################### [henning@archlinux ~]$ sudo rm -rf test/ [henning@archlinux ~]$ mkdir test [henning@archlinux ~]$ mkarchroot test/root base-devel ==> Creating install root at /home/henning/test/root ==> Installing packages to /home/henning/test/root :: Synchronizing package databases... error: could not open file /home/henning/test/root/var/lib/pacman/sync/download-AIUAX3/core.db.part: Permission denied error: failed to setup a download payload for core.db error: failed to synchronize all databases (failed to retrieve some files) ==> ERROR: Failed to install packages to new root ==> ERROR: Failed to install all packages

2 1

dbus-broker-units provides dbus-units
by u34＠net9.cf 25 Sep '24

25 Sep '24

$ pacman -Qi dbus-broker-units | grep Provides Provides : dbus-units Seeing that, I assumed it is enough to have the dbus-broker-units package, without having the dbus-units package. That assumption turns to be wrong. While trying to build a chroot environment, if the dbus-units package can not be obtained, the process fails. :: Retrieving packages... dbus-units-36-4-x86_64.pkg.tar.zst failed to download error: failed retrieving file 'dbus-units-36-4-x86_64.pkg.tar.zst' from the_repo : The requested URL returned error: 404 warning: failed to retrieve some files error: failed to commit transaction (failed to retrieve some files) Errors occurred, no packages were upgraded. ==> ERROR: Failed to install packages to new root ==> ERROR: Failed to install all packages One reason for the explicit search for dbus-units is that systemd explicitly depends on dbus-units. $ pacman -Qi systemd | grep dbus-units Depends On : ... dbus dbus-units ... Am I confused? Why the chroot must have a dbus-units package? -- u34

2 1

mkarchroot and Manual intervention for pacman 7.0.0 and local repositories required
by u34＠net9.cf 24 Sep '24

24 Sep '24

$ pacman -Qo mkarchroot /usr/bin/mkarchroot is owned by devtools 1:1.2.1-1 Referring to https://archlinux.org/news/manual-intervention-for-pacman-700-and-local-rep… and to https://wiki.archlinux.org/title/DeveloperWiki:Building_in_a_clean_chroot#S… with a modified pacman.conf: error: could not open file /home/user/chroot/root/var/lib/pacman/sync/download-o6cp5z/core.db.part: Permission denied error: failed to setup a download payload for core.db error: failed to synchronize all databases (failed to retrieve some files) ==> ERROR: Failed to install packages to new root ==> ERROR: Failed to install all packages The modified pacman.conf is using links created with ln -s /var/lib/pacman/sync/*.db /var/cache/pacman/pkg/ I have also tried the naive approach of --- a/usr/bin/mkarchroot 2024-06-23 21:00:54.158267054 +0000 +++ b/usr/bin/mkarchroot 2024-09-22 19:31:52.984269364 +0000 @@ -73,6 +73,7 @@ [[ -e $working_dir ]] && die "Working directory '%s' already exists" "$working_dir" mkdir -p "$working_dir" +chown :alpm "$working_dir" lock 9 "${working_dir}.lock" "Locking chroot" In addition, the most basic suggestion of https://wiki.archlinux.org/title/DeveloperWiki:Building_in_a_clean_chroot#S… , namely # mkarchroot /home/user/chroot/root base-devel also fails. Though it could be due to a usually working XferCommand in pacman.conf. Does mkarchroot broken? -- u34

2 5

spam on the forum
by funny0facer 22 Sep '24

22 Sep '24

Could someone please ban https://bbs.archlinux.org/profile.php?id=162864 on the forum? This person is posting pictures, I really do not want to see. Thanks

3 5

pacman 7 - why is #ParallelDownloads = 5 commented by default now?
by David C. Rankin 21 Sep '24

21 Sep '24

Devs, As I merged .pacnew changes for pacman.conf, I notice the parallel download option is now commented by default. I left it that way. But why is that now the default after having parallel transfer come in as a new feature a year or two ago? -- David C. Rankin, J.D.,P.E.

2 2

Re: Arch Linux public upload server
by Rein Fernhout (Levitating) 19 Sep '24

19 Sep '24

I sent this message yesterday but I accidentally forgot to add the mailing list as a recipient. -------- Original Message -------- Subject: Re: Arch Linux public upload server Date: 2024-09-04 14:28 From: "Rein Fernhout (Levitating)" <me(a)levitati.ng> To: Robin Candau <antiz(a)archlinux.org> On 2024-09-04 09:56, Robin Candau wrote: > On 9/4/24 1:04 AM, Polarian wrote: >> Hello, >> > > Hello, > Hello, >> >> So to spur such a discussion, why should it be rustypaste? >> What makes it better than the competition? >> > > I think you're taking Orhun's mail/proposition the wrong way. > > First of all, Arch Linux and rustypaste are not corporate so there's no > "corporate buddies"/shareholders in that context. I personally have not > been bothered by any potential "conflict of interest" in their > suggestion. > > Secondly, Orhun's demand is not "Should Arch Linux host a public > pastebin/upload server (and, if so, could it be rustypaste)?" but > "rustypaste is looking for a public instance [2], could Arch Linux host > it?" > > Here again, I feel like this an important nuance to understand why the > discussion is specifically about rustypaste (and not about any other > similar services). > > > [1] > https://lists.archlinux.org/archives/list/arch-general@lists.archlinux.org/… > [2] https://github.com/orhun/rustypaste/issues/326 Though I do think that if we were to have a pasting service, I would like to use a battle-proven tool with proper moderation capabilities like 0x0. But in generally I think we just shouldn't start. The amount of abuse pasting services see is absolutely staggering. The only way I think we could manage is, like mentioned before, by protecting it with our SSO account authentication. But rustypaste doesn't rely on a frontend, it works similarly to 0x0. So what does it add to our community? The IRC already recommends use of 0x0 or termbin.com. Both of which rely on open source software that can be selfhosted by users with ease. I would maybe see added value if our pastebin service was graphical, as this would make it more accessible. Authenticated users could submit pastes and anyone could view them. But rustypaste isn't graphical. So I do a agree with Polarian that we should first consider "do we want a pasting service" and then "do we want rustypaste". Even if that was not Orhun's original question.

4 4

[arch-general] Is wine-mono now integrated into wine (9.17)?
by Javier 18 Sep '24

18 Sep '24

Hello ! I'm guessing wine-gecko is now integrated into wine, since it recently got dropped from the repos but it didn't even make it to AUR. Is that the case? Thanks ! -- Javier

2 2

Arch Websites vs Dark Mode
by Genes Lists 16 Sep '24

16 Sep '24

I see that the wiki and gitlab both have dark modes but I didn't see how to activate dark mode for any other part of the website. Did I miss where to turn this on or if it's not yet available are there any plans to add it? Not sure how others feel, but might be nice to offer it. thanks! -- Gene

4 3

Guest issue after virtualbox upgrade from 7.0.20-1 to 7.1.0-1
by Ralf Mardorf 15 Sep '24

15 Sep '24

Hi, the execution of a Windows 11 (64-bit) guest works, with the exception that the shutdown does not end, but this was already a problem under 7.0.20-1, which you can live with. Under 7.0.20-1, a Windows 10 (64-bit) guest, including the shutdown, worked perfectly smoothly. Unfortunately, the Windows 10 startup freezes under 7.1.0-1. Its log ends with "00:01:40.881013 EHCI: USB Suspended". By the manager settings I changed a few settings, to use mainly the same settings for the Windows 10 guest, that are used by the Windows 11 guest. In addition I migrated from USB 2.0 to USB 1.1 for the Windows 10 guest. Now USB is operational and I couldn't see any useful information by the log file. I'm in a hurry as I have to go to work and will read the logs again later at my leisure. Has anyone had similar experiences? FWIW pacman.log mentions an issue: [2024-09-13T05:02:52+0200] [ALPM] upgraded virtualbox (7.0.20-1 -> 7.1.0-1) [2024-09-13T05:02:52+0200] [ALPM-SCRIPTLET] VBoxManage: error: Failed to create the VirtualBox object! [2024-09-13T05:02:52+0200] [ALPM-SCRIPTLET] VBoxManage: error: Most likely, the VirtualBox COM server is not running or failed to start. After restarting the host trying to start the VirtualBox manager GUI for the first time also failed with the object error, but after that it always started without an issue. extra/virtualbox 7.1.0-1 extra/virtualbox-guest-iso 7.1.0-1 extra/virtualbox-host-dkms 7.1.0-1 chaotic-aur/virtualbox-ext-oracle 7.1.0-1 $ uname -r; cat /proc/cmdline 6.10.9-arch1-2 BOOT_IMAGE=/boot/vmlinuz-linux root=/dev/disk/by-label/m1.archlinux ro threadirqs ibt=off Regards, Ralf

1 1

libreoffice - new versioning download frequency - multiple updates per-week?
by David C. Rankin 12 Sep '24

12 Sep '24

Devs, With the advent of the new version scheme for libreoffice, there is a dramatic increase in the frequency of package updates lately, sometimes a new version is packaged daily or every other day, e.g. [2024-08-26T19:24:24-0500] [ALPM] upgraded libreoffice-still (24.2.5-1 -> 24.2.5-2) [2024-09-02T23:01:03-0500] [ALPM] upgraded libreoffice-still (24.2.5-2 -> 24.2.5-3) [2024-09-07T01:54:56-0500] [ALPM] upgraded libreoffice-still (24.2.5-3 -> 24.2.6-1) [2024-09-09T16:53:22-0500] [ALPM] upgraded libreoffice-still (24.2.6-1 -> 24.2.6-2) [2024-09-10T18:48:31-0500] [ALPM] upgraded libreoffice-still (24.2.6-2 -> 24.2.6-3) Granted with gcc changes, etc.. many rebuilds may be needed as those changes are incorporated, and it may just be temporary, but libreoffice is a 150M+ package. I can imagine that adds up from a bandwidth standpoint for Arch and it's mirrors. Is the libreoffice new package frequency just a temporary thing, or is this here to stay? More a curiosity than anything else. I can look back nearly a decade in the pacman.log and it wasn't uncommon to have months when there was no libreoffice update, and rarely 3 or more a month. -- David C. Rankin, J.D.,P.E.

1 0