- Arch-general - lists.archlinux.org

archlinux-keyring-wkd-sync.service: Errors refreshing keys
by Ariadna Vigo 02 Dec '25

02 Dec '25

Hi everyone, For a couple of weeks now, I have been noticing that archlinux-keyring-wkd-sync.service fails every single time it is fired up by its accompanying timer. Restarts are attempted, but fail nonetheless. The reason why is that many keys are not retrievable via WKD, I imagine because they are not set up to be so by their respective owners. I have waited enough to confirm that it wasn't some network error on my side before sending this message. Journal info on the failing keys is available at [1]. All other keys are correctly retrieved via WKD and refreshed if that's the case, so this isn't a critical issue. However, if some key isn't available via WKD, maybe should it be silently dropped, so that errors are reserved to something bad happening to keys that are available via the protocol? Keys that aren't available via WKD will get updated via regular updates to the archlinux-keyring package in any case. The thing is that I find having a service failing this way (with a timer that is enabled by default) avoidable. That 'systemctl status' shows a degraded status every week because of this false positive makes it harder to detect when something else might have gone wrong that could require proper attention. Best, Ariadna [1]: https://paste.sr.ht/~ariadna/0f509b0c5823ab225523c989617967820db07a3a -- Ariadna Vigo https://ariadnavigo.xyz gpg 0xC948873069856D6D

6 11

FLTK 1.4.4 package in extra repo
by Robert Arkiletian 30 Nov '25

30 Nov '25

Hello, I am one of the upstream maintainers of pyfltk https://pyfltk.sourceforge.io/ I adopted the orphaned python-pyfltk package in the AUR. In the process of creating a PKGBUILD, I noticed that the FLTK 1.4.4 package in the extra repo is not built with Wayland support. One of the main new features of FLTK 1.4 is that it now has native Wayland support. Hence, pyfltk 1.4 now also has native Wayland support. I would like to request that FLTK be re-built with Wayland support. Here are some of the missing makedepends packages cairo pango libdecor wayland-protocols There may be more packages required. You can verify that wayland is supported from the output of the cmake command. The README.Wayland.txt in the FLTK source tree describes the dependencies for other distros like Debian and Fedora.

3 4

When using wpa_supplicant, TDLS connection attempts break Wi-Fi on some chipsets
by Liliane Fontenot 28 Nov '25

28 Nov '25

Hi all! Happy holidays to anyone celebrating today. I'm writing about an issue I found with TDLS breaking my Wi-Fi connection on a laptop MediaTek modem (MT7921). This seems to be reported in numerous other places, although most users didn't manage to track the issue all the way to its source. I'm hoping that other people can confirm affected models and help me put together an actionable bug report and figure out where to send it (either to wpa_supplicant or to specific kernel driver mailing lists). TDLS is a Wi-Fi protocol allowing the creation of direct connections between two Wi-Fi clients on the same local network segment, without disrupting the existing connection. [1] Support for TDLS in wpa_supplicant has been enabled by default since version 2.9 [2] and Arch Linux apparently enabled it before that. Applications like KDE Connect that transfer data between an Android device and a PC can cause these TDLS connections to be initiated. A quick search of the KDE Connect code base turned up nothing for TDLS, and so I think wpa_supplicant is creating these connections automatically in response to large transfers between Wi-Fi clients. (Alternatively, perhaps the Android networking stack detects support for TDLS and tries to create the connection.) KDE Connect therefore frequently turns up in reports of this issue, although I have also seen Syncthing and even ssh mentioned as well. For me, the problem can reliably be reproduced in the following way: when connected (via NetworkManager + wpa_supplicant) to the same access point as my phone, I try to send a file from my phone (Android) to my computer using KDE Connect. This will generally work. Additional files, however, will transfer very slowly or fail entirely. Around the same time, the Internet connection on my computer will suddenly become slower (but not break entirely, which is why it took me a long time to figure out the problem). The source of the issue seems to be this: after wpa_supplicant tries to initiate a TDLS connection with my phone, the rate of dropped packets between my computer and my access point goes up, disrupting and slowing my connections (and usually breaking KDE Connect itself). Some users have even reported that they are de-authed from their access point, resetting the connection (which has the side effect of fixing the issue with dropped packets). [3] This causes the following log entry to appear: Jun 21 15:54:09 wpa_supplicant[2470]: TDLS: Creating peer entry for 0e:4d:96:35:0e:38 Jun 21 15:54:09 wpa_supplicant[2470]: TDLS: Dialog Token in TPK M1 1 Jun 21 15:54:09 wpa_supplicant[2470]: nl80211: kernel reports: key addition failed Jun 21 15:54:09 wpa_supplicant[2470]: TDLS: Failed to set TPK to the driver Many users have reported effectively this same issue where TDLS breaks their Internet connection, and they almost all have this same log entry. Here is a sampling: [4] [5] [6] [7] [8] [9] [10]. Some users resolved their issue by switching the NetworkManager backend to iwd. I confirmed that this fixes my issue too. I believe, however, this the reason this works is probably that iwd doesn't support the TDLS protocol. There's a bug here, and the ideal resolution would be to get a fix for it either in wpa_supplicant or in the individual WiFi drivers that may have issues. I was also able to resolve the issue by simply rebuilding the latest Arch Linux wpa_supplicant package while disabling the CONFIG_TDLS build option. My problem at this point is that I'm not sure how to get any further information that might help developers find and resolve the bug. Any help with this would be appreciated! Warmly, Liliane [1] https://en.wikipedia.org/wiki/TDLS [2] https://git.w1.fi/cgit/hostap/commit/?id=39f29f250a [3] https://www.reddit.com/r/kde/comments/1lha5x8/cant_send_files_using_kde_con… [4] https://bbs.archlinux.org/viewtopic.php?id=276748 [5] https://bbs.archlinux.org/viewtopic.php?id=283498 [6] https://www.reddit.com/r/linuxquestions/comments/12ge74g/uploading_big_file… [7] https://discussion.fedoraproject.org/t/laptop-freezes-after-updated-to-kern… [8] https://discussion.fedoraproject.org/t/network-gets-cut-off-when-connecting… [9] https://bbs.archlinux.org/viewtopic.php?id=256314 [10] https://github.com/GSConnect/gnome-shell-extension-gsconnect/issues/530

2 1

thunderbid error in SmtpServer.sys.mjs when sending messages
by Łukasz Michalski 25 Nov '25

25 Nov '25

Hi, My Thunderbird instances suddenly started to choke on sending emails to smpt.office365.com (not sure for others). Both my 139.0.2-1 and 143.0.1-2 instances have the same problem: 1. I click 'send' and TH is stuck on "Sending message" dialog box. 2. devtools shows error in console: 08:57:41.360 NS_ERROR_FILE_NOT_FOUND: Component returned failure code: 0x80520012 (NS_ERROR_FILE_NOT_FOUND) [nsIFile.fileSize] SmtpServer.sys.mjs:598 3. When I click 'Cancel' in dialog box from (1) then, 'Send' again the problem repeats. 4. When I wait for TH to give up and click 'Send' again message is sent successfully. Anyone sees similar behavior? Regards, Łukasz

2 2

Cannot always resume from hibernate on linux 6.17.*
by gesh 25 Nov '25

25 Nov '25

Hi, For an unclear reason, sometimes when resuming from hibernate on a system running linux and xorg, the screen turns black and unresponsive to even sysrq. Hibernation from a non-xorg vty works well, but this is insufficient for my uses. Moreover, sometimes even when resuming from hibernation does work, within a couple minutes of the resume, the screen flashes with a blue QR code for maybe half a second (which, if this is meant to give me an opportunity to scan it, is woefully inadequate), then turns black and unresponsive. In contrast, with linux-lts, I'm so far not having any similar problems. In particular, from initial searches I was under the impression that the problem was in one of the graphics drivers, and tried switching from nvidia to nouveau and finally disabling both, but that didn't influence the results. This situation has been plaguing me for about two months now -- the difficulty diagnosing the issue has prevented me from dedicating the time to investigate. Does anybody know at least of where I can look to try to get a better picture of what's going on? Thanks, Gesh The software versions for the relevant parts of the stack (haven't tested linux-lts+nvidia yet, but the version listed is the one that broke with linux) linux 6.17.8.arch1-1 linux-lts 6.12.59-1 mesa 1:25.2.7-1 xorg-server 21.1.20-1 i3-wm 4.24-1

3 2

AUR - connection issues 11/17 at 0600 zulu - hopefully just maintenance?
by David C Rankin 17 Nov '25

17 Nov '25

All, Last night around 0600 GMT, I had difficulties reaching AUR. I would get "nginx gateway 502" errors and "Command Not Found" rendered in the browser at other times. When connections were successful, they would take 10 - 20 seconds to actually return content. Then clicking links on a package page, like "View PKGBUILD" would fail (as described above) Hopefully this was just maintenance and not another surge of bots. I thought I would drop a message about it -- though I have every confidence the web wizards at Arch already know. -- David C. Rankin, J.D.,P.E.

1 0

Fwd: How do I advise users to install the arch package for Cpdf?
by Andy Pieters 17 Nov '25

17 Nov '25

On Mon, 17 Nov 2025 at 15:56, Ralph Corderoy <ralph(a)inputplus.co.uk> wrote: > Hi John, > > Max wrote: > > "cpdf can be installed from the AUR > > (https://aur.archlinux.org/packages/cpdf) using the normal methods" > > He's right, an ArchLinux user will know about the AUR, but you can refer > them to https://wiki.archlinux.org/title/Arch_User_Repository as > a starting point if you wish to make sure. > > I think it's important to keep the disconnect and do point people to the wiki like Max says rather than just say "install helper", *and* to highlight that AUR is *user* content, not subject to the normal QA and vetting from the main repository, and not endorsed by Arch

1 0

How do I advise users to install the arch package for Cpdf?
by John Whitington 17 Nov '25

17 Nov '25

Hello, Someone has kindly provided an Arch package for a piece of software I wrote: https://aur.archlinux.org/packages/cpdf I had thought it could be installed with pacman -S cpdf, and not being an arch user, that's what I'd put in the documentation. In fact, this is not true, because it's in the AUR. I installed Arch, and yay, and ran: yay -S cpdf which completed successfully. But running 'cpdf' at the command line cannot find the command. And yet, from the comments, other people are using it successfully. Any advice on what I can tell my users? Or a better venue to ask this question? Thanks, John -- John Whitington Director, Coherent Graphics Ltd https://www.coherentpdf.com/

3 2

Connection timeout when accessing Arch GitLab
by DodoGTA GT 16 Nov '25

16 Nov '25

Hello I've been having issues accessing the Arch Linux GitLab (gitlab.archlinux.org) for more than a week now Every time I try to access that domain, the connection keeps timing out (and `ping` command shows a 100% packet loss) Here's what `tracepath` command outputs for reference: ❯ tracepath gitlab.archlinux.org 1?: [LOCALHOST] pmtu 1500 1: _gateway 0.818ms 1: _gateway 0.674ms 2: <ISP address> 9.205ms 3: decix2-gw.hetzner.com 30.104ms 4: core21.fsn1.hetzner.com 34.806ms 5: ex9k2.dc13.fsn1.hetzner.com 34.676ms 6: no reply What's strange is that I can access every other Arch Linux service fine (including the AUR which has been unstable in the past few weeks) Also I've already set up DNS-over-HTTPS (so DNS shouldn't be an issue) and I've had an Arch DevOps person check my IP to confirm it isn't blocked (so it might be some further upstream issue like Hetzner?) And finally, accessing the website via Tor/Tor Browser works and I was even able to do a `git push` (so that's a workaround I could use for now) Hopefully a resolution can happen in the near future (because I'm a bit annoyed by this issue at this point) Thanks

2 2

Re: Possible to script pacman -U to automate swap between dovecot 2.3/2.4, or 2.4/2.3?
by Pocket 07 Nov '25

07 Nov '25

On 11/7/25 3:20 PM, David C Rankin wrote: > On 11/6/25 5:01 PM, Pocket wrote: >> I use the btrfs filesystem, I create a subvolume for version 1 and >> install the base packages and the "old version" It is configured and >> boots and "runs" using systemd-nspawn. >> >> I then create a subvolume for version 2, install the base packages and >> the "old version" It is configured and boots and "runs" using >> systemd- nspawn. >> >> I then can use machinectl to run/start a "version" and then I can shut >> it down and run the other, ie >> machinectl start "version 1" >> machine login "version 1" >> >> when I am finished then >> >> machinectl start "version 2" >> machine login "version 2" >> >> systemd containers are your friend > > Thanks! > > That's something I've not looked into yet. I've avoided btrfs and > stayed on ext4 --- I also run opensuse and lived through the development > years when btrfs would regularly exhaust drive space leaving users > without a usable system. > > It's matured, but there are also distros now dropping it, so that has > been another ding against my trying it. The systemd container aspect > does as a huge capability to its arsenal. I'll have to read up. > ls /var/lib/machines/ DNS PiHole machinectl list MACHINE CLASS SERVICE OS VERSION ADDRESSES DNS container systemd-nspawn archarm - - PiHole container systemd-nspawn archarm - - - base packages are installed to DNS and PiHole. bind is installed into the DNS subvolume and pihole is installed into the PiHole subvolume. They are started at boot. They are separate from the "Root" subvolume that the system boots to. This keeps all the "services" separate from the boot system and I don't have the bind and pihole packages installed/mixed into the "root system". I am currently creating a subvolume for nginx to provide web service to the local network. btrfs subvolume list /mnt ID 256 gen 9111 top level 5 path Root ID 257 gen 9099 top level 5 path Home ID 258 gen 8452 top level 5 path Pkgs ID 259 gen 263 top level 256 path Root/var/lib/portables ID 260 gen 9100 top level 256 path Root/var/lib/machines ID 261 gen 9111 top level 5 path PiHole ID 262 gen 263 top level 261 path PiHole/var/lib/portables ID 263 gen 1005 top level 261 path PiHole/var/lib/machines ID 264 gen 9110 top level 5 path DNS ID 265 gen 1005 top level 264 path DNS/var/lib/portables ID 266 gen 1005 top level 264 path DNS/var/lib/machines Overall: Device size: 27.31GiB Device allocated: 4.57GiB Device unallocated: 22.74GiB Device missing: 0.00B Device slack: 0.00B Used: 2.35GiB Free (estimated): 24.65GiB (min: 13.28GiB) Free (statfs, df): 24.65GiB Data ratio: 1.00 Metadata ratio: 2.00 Global reserve: 7.53MiB (used: 0.00B) Multiple profiles: no Data,single: Size:4.01GiB, Used:2.10GiB (52.32%) /dev/mmcblk0p2 4.01GiB Metadata,DUP: Size:256.00MiB, Used:129.83MiB (50.71%) /dev/mmcblk0p2 512.00MiB System,DUP: Size:32.00MiB, Used:16.00KiB (0.05%) /dev/mmcblk0p2 64.00MiB Unallocated: /dev/mmcblk0p2 22.74GiB I have compress=lzo turned on all volumes -- Hindi madali ang maging ako

1 0