- Arch-general - lists.archlinux.org

Is there a way to submit a one-shot update to an Aur package without becoming the maintainer?
by David C. Rankin 03 Nov '24

03 Nov '24

Arch and Aur Folks, I updated https://aur.archlinux.org/packages/mongodb44 to build with the last 4.4.29 source and patched for boost 1.86. The package builds fine now against current Arch. I wanted to pass along the results of those efforts and update the Aur package, but didn't want to take on the maintainer role. Is there a way to provide an update for the package and bring in back into date without becoming the maintainer? I have a signed source package I'm happy to pass along if that is something the admins can do. There is about 8 hours of effort to that went to find and fix all the issues, I'd like to make that available to whoever may need the 4.4 version of mongodb. I didn't think I'd ever need it, but it is the last version that wasn't built requiring avx - so it is useful if you need it for those cases or need it in a VM, etc.. (note: some parts of a 3rd party package included along with the source will need a significant 'type' change to suppress the char8_t -Wc++20-compat type warning, but that's just something for the future, and there is one more -Wredundant-move I need to fix in src/mongo/db/matcher/expression_array.h that somehow got missed in the hundreds of thousands of lines of output) I've tried posting to aur-general, but my posts do not ever make it to the list. I don't mind picking up the maintainer role, but am hesitant to get caught be another large boost deprecation/removal during a busy week and not have time to devote to it. Is a one-shot update something that can be done for this package without becoming the maintainer? -- David C. Rankin, J.D.,P.E.

3 8

Status of package superslicer
by Björn Försterling 30 Oct '24

30 Oct '24

Hi, was the package "superslicer" removed from pacman? $ sudo pacman -Ss superslicer $ echo $? 1 Also I cannot find the package "slic3r" as well. The only slicer that pacman supports now seems to be the "prusa- slicer". Does anyone know why these packages where removed? Regards Björn

1 0

Request for review of two merge requests (Fcitx5-Mozc, Fcitx-Mozc)
by Masato TOYOSHIMA 29 Oct '24

29 Oct '24

Hello everyone I've been using Arch Linux in containers and have tried several Arch-based desktops. I'd like to discuss two Japanese input packages that haven't been updated for about two years. I've created merge requests for major upgrades for both packages. Among Japanese input packages, Mozc likely has the most users currently. There are two Mozc packages in the Arch Linux repository: Fcitx5-Mozc and Fcitx-Mozc. While Fcitx5 has become the main development focus, Fcitx-mozc is still packaged in newer distributions. These two packages haven't been updated for about two years. I created merge requests for major upgrades about two weeks ago. I've tried contacting the previous package maintainers and those assigned by bots, but there's been no response. I'm posting here hoping to catch the attention of those with merge permissions. I'd appreciate if you could review the merge requests for these two packages when you have time. As additional information, the original google/mozc has moved to Bazel build for the Linux version. Fcitx5 already supported Bazel build, and Fcitx (fcitx4) has recently added support for it as well. Therefore, both packages can now be built with Bazel without additional patches or PRs. I've been updating the merge requests to keep them aligned with the latest commits. Now that Fcitx(4)'s Bazel support is complete, it seems like a good time for review. Please review the merge requests when possible. I apologize for the long message. Thank you for your consideration. https://gitlab.archlinux.org/archlinux/packaging/packages/fcitx5-mozc/-/mer… https://gitlab.archlinux.org/archlinux/packaging/packages/fcitx-mozc/-/merg…

1 0

Nextcloud - Archlinux packaged apps verse Nextcloud store - What is the difference?
by David C. Rankin 23 Oct '24

23 Oct '24

Devs, I've used nextcloud in Arch for a couple of years. When I installed, it was just the Nextcloud webapp that was available and the applications were packaged by nextcloud with additional apps available in the nextcloud store. Now there are "nextcloud-app-xxxx" packages provided by Archlinux, e.g. nextcloud-app-bookmarks nextcloud-app-calendar nextcloud-app-contacts nextcloud-app-main nextcloud-app-news nextcloud-app-notes nextcloud-app-spreed Did something change that now means I should be installing the nextcloud-app-xxxx packages instead of the apps that are already in Nextcloud? The wiki https://wiki.archlinux.org/title/Nextcloud is silent on the packaged apps. Are they different from what is available in the nextcloud store? Better? Just trying to figure out why they are packaged and if I should switch to them. -- David C. Rankin, J.D.,P.E.

2 3

About AUR accreditation policy
by Óscar García Amor 22 Oct '24

22 Oct '24

Hello foks, I tell you, about 8 hours ago a package that Radu C. Martin and I were maintaining (textpieces) was moved from the AUR to extra. In this movement (which seems perfect to us, every package officially maintained is welcome and better than this in the AUR) our names have disappeared[1]. I don't want you to misunderstand me, I'm not crying foul about this, but it doesn't seem fair to me that the person (or persons) who kept the package for a while until it became official is not credited. If we in the AUR keep the names of those who preceded us in the maintenance of the package, I understand that when a package moves from the AUR to an official repository the same should happen. This happens when a package is dropped to AUR. I don't know if there is a directive for this, but if not, maybe it's time to write one. And as I said, I don't mind that my name doesn't appear as a contributor right now (I keep my AUR packages for myself and for people to benefit from it, not for fame or similar nonsense), but I do find it a bit ugly that the effort involved in doing that work is not recognized, and not by me, but by everyone. Greetings. PS: I would not want accusations or ugly things to come out of this. I just want to note that there should be a policy that the AUR maintainer's work should be credited when a package is moved to an official repository. [1]: https://gitlab.archlinux.org/archlinux/packaging/packages/textpieces/-/blob… -- Óscar García Amor | ogarcia at moire.org | http://ogarcia.me

13 18

Arch Linux stuck on legacy iptables by default
by Martin Rys 18 Oct '24

18 Oct '24

Arch Linux still uses the legacy iptables backend, as opposed to the nft backend that every other distribution uses. This has been reported since 2021 with no response: https://gitlab.archlinux.org/archlinux/packaging/packages/iptables/-/issues… To me it looks like a direct, simple migration, and it just needs someone with the needed perms to look at it, but I could be wrong. It's assigned to the notoriously overburdened Felix, is there anything a non-core member can do, other than wait? It's not a *large* issue, it's just mildly annoying to have workarounds in my deploys, as it is installed on any system due to iproute2 needing an iptables backend, and some things just outright don't work with that, requiring the nft one. Thanks, Martin

6 9

Network Issues After Pacman Upgrade
by Björn Försterling 14 Oct '24

14 Oct '24

Hello, I just ran a "pacman -Syu" and am having issues with the internet connection. I have around 10 seconds with a working connection to the internet followed by 10 seconds without being able to establish a connection. For 10 seconds I am able to curl google.com or telnet its port 443, and then I get blocked for 10 seconds. There was an update to "networkmanager": [2024-10-13T15:39:41+0200] [ALPM] upgraded networkmanager (1.48.10-1 -> 1.50.0-1) Could this be the cause? Does anyone else experience similar problems? I had the same problem a few weeks ago. Regards Björn

5 5

Strange config /etc/profile.d/utf8.sh added on recent upgrade
by Tadeas Uhlir 13 Oct '24

13 Oct '24

Hey all, Today, I noticed that my alias for "ls" is not working as usual. I alias "ll" to "LC_COLLATE=C ls -alF" to have a long listing with hidden dirs at the top. However, I noticed that the sorting doesn't work all of a sudden. After some digging around I found out that this is due to "LC_ALL" being set to "en_US.utf8". This variable is being set from a file at "/etc/profile.d/utf8.sh". Two notes about this: 1. For the life of me I cannot figure out which package upgrade/install created this file. Just going by a pacman log the file creation time corresponds to when "pacman --noconfirm -S autoconf libtool m4 automake" ran. But those packages don't seem to have much to do with locale. And googling the path of this file doesn't seem to yield much, clearly it's not a very common config file across linux distros. 2. Once I find the culprit, I would love to know what is the rationale behind explicitly setting "LC_ALL". Even the archwiki mentions that this should only be done during tests: https://wiki.archlinux.org/title/Locale#LC_ALL:_troubleshooting . And, indeed, seems like a bad idea given that it overrides all LC_ vars. -- Best Regards, Tadeáš Uhlíř

2 2

Periodic aged package report
by Genes Lists 11 Oct '24

11 Oct '24

Periodic check of Arch out of date packages. This considered are purely my opinion of those that are sufficiently 'relevant'. Apologies to any others. The really good news is that there really aren't many ! Big congrats for the hard work keeping Arch awesome and up to date. Note that, some are excluded when they are work in progress and/or involve a significant amount of work (provided they are not "too" old, which none are). Arch is doing an amazing job keeping our core/key packages like compiler tools, python, systemd, kernel and other vital packages very up to date. What remains are a few that are relevant enough (imho) to make the list. I am using 3 metrics of age: - Num-B The number of releases we are behind Some packages have frequent updates some do not. - Age The number of days between the arch version and the latest - New: How recent was the latest released (days ago from today) Package Arch Curr Date num-b Age New ------------------ ---------- ------ -------- ------ ------ ---- apparmor 3.1.7-5 4.0.3 08/15/24 4 195 55 biber 1:2.19-4 2.20 03/21/24 1 321 202 dkms 3.0.12-1 3.1.0 10/01/24 2 373 8 python-cryptography 42.0.8-1 43.0.1 09/03/24 2 91 36 python-setuptools 1:69.5.1-1 75.1.0 09/13/24 22 153 26 yubioath-desktop 5.1.0-3 7.1.0 09/25/24 12 1234 14 (1) (1) Now called yubico-authenticator -- Gene

5 7

Valgrind problem
by Frank 07 Oct '24

07 Oct '24

Good evening, I'm trying to debug an issue with Evolution and the Gnome guys asked me to use valgrind. However, when running this command I get an error I'm stuck with: valgrind --num-callers=30 --leak-check=no --track-origins=yes /usr/lib/evolution-calendar-factory -w ==4915== Memcheck, a memory error detector ==4915== Copyright (C) 2002-2024, and GNU GPL'd, by Julian Seward et al. ==4915== Using Valgrind-3.23.0 and LibVEX; rerun with -h for copyright info ==4915== Command: /usr/lib/evolution-calendar-factory -w ==4915== valgrind: Fatal error at startup: a function redirection valgrind: which is mandatory for this platform-tool combination valgrind: cannot be set up. Details of the redirection are: valgrind: valgrind: A must-be-redirected function valgrind: whose name matches the pattern: strcmp valgrind: in an object with soname matching: ld-linux-x86-64.so.2 valgrind: was not found whilst processing valgrind: symbols from the object with soname: ld-linux-x86-64.so.2 valgrind: valgrind: Possible fixes: (1, short term): install glibc's debuginfo valgrind: package on this machine. (2, longer term): ask the packagers valgrind: for your Linux distribution to please in future ship a non- valgrind: stripped ld.so (or whatever the dynamic linker .so is called) valgrind: that exports the above-named function using the standard valgrind: calling conventions for this platform. The package you need valgrind: to install for fix (1) is called valgrind: valgrind: On Debian, Ubuntu: libc6-dbg valgrind: On SuSE, openSuSE, Fedora, RHEL: glibc-debuginfo valgrind: valgrind: Note that if you are debugging a 32 bit process on a valgrind: 64 bit system, you will need a corresponding 32 bit debuginfo valgrind: package (e.g. libc6-dbg:i386). valgrind: valgrind: Cannot continue -- exiting now. Sorry. THX Frank

3 4