- Arch-general - lists.archlinux.org

Keeping desktop PC and laptop in sync
by Joel 07 Nov '22

07 Nov '22

Hi, as we have a discussion ongoing about setting up a new PC, I thought I could jump in with my current issue : I have retired my laptop and bought a desktop PC. Gone through the fresh installation process (once every 12 years is OK to declutter a bit), and rsync'ed my home directory (parts of it, mostlyskipped ~/.local and~/.config where I found some 10-20 years old files from programs I used once but never since). Now I still want to use my laptop when I'm on the road, which happens rather seldomly (once every 2 weeks), i.e. have my files, my bookmarks etc available on my laptop. Also not only have to sync my /home/$USER but also /root and some more mountpoints (source code, website stuff, audio, video, pictures), totaling around 1.5TB of data I've been playing around with rsync and adding many many exclude rules to a) limit the amount of files being transferred (eg exclude ~/.cache, browser's cache, thunderbird's imap cache etc) so I can keep an overview if something is getting overwritten that I don't want b) keep certain files (especially XFCE config etc) distinct as eg I have 2 monitors on my PC vs only 1 on the laptop -> different menus, window settings etc Now this is pretty tedious, initial rsync's dry-run showed ~700k files that would be sync'ed, which I brought down to 1k after several iterations and adjustment of exclude settings, and this is so far my first run - currently using my laptop and will sync back when I'm home again... let's see how many files will differ then ! So I'm looking for more easy-to-use alternatives. Most important thing : I want dry-run functionality to prevent unwanted data loss, which means automatic sync tools are not the way to go ! I think synthing is such a candidate, unless I missed the dry-run option somewhere. Have used unison many many years ago, but I think it was also not perfect and had issues handling thousands/millions of files. Heard of some new systemd homesync (?), but didn't have a look at it now, sounds also like it automatically syncs. Any thoughts / recommendations / advices ? Many thanks Joel

6 8

Re: [arch-dev-public] openssl 3.0
by Genes Lists 07 Nov '22

07 Nov '22

Just a note to thank you all for pushing this over the line. Obviously a huge effort and very big thanks!! Everything I've tested so far works just fine : dovecot, postfix, ldns, rust/cargo, nginx-mainline, cyrus-sasl, nsd, unbound, cryptsetup, stunnel and kernel signed modules. Looking forward to testing python 3.11 as well!! Thank you for all the work you do keeping things running so well. gene

5 5

kea_config tool available
by Genes Lists 05 Nov '22

05 Nov '22

In the spirit of continuing to give back to the community, I'm happy to share my kea configuration tool. kea_config is available in the AUR [1] with source on github [2] kea is the current ISC dhcp server software taht offers a nice feature set, including a hot standby to pick up should the primary be unavailable. Its power also lurks behind a somewhat complicated configuration suite that, at least for me, isn't the most human friendly and creates a burden to keep things appropriately consistent. kea_config provides a nice and clean way to generate the required config files from a simple, single config. It generates the set of configs for each of primary, backup and standby servers which are then copied to each server; each can then use a link to primary or standby or backup config as appropriate. If you only run primary, that's fine too :) For host reservations it uses hostnames and looks up the IP - thus further helping reduce risk of human error. Hope others find it helpful. regards, gene [1] https://aur.archlinux.org/packages/kea_config [2] https://github.com/gene-git/kea_config

1 0

Setting up new computer to be identical to old one
by Spencer Collyer 01 Nov '22

01 Nov '22

I am looking to replace my current desktop PC, which is getting on for 8 years old now, with a more up-to-date one. Ideally I want the new one to be set up the same as my current one, which has been running on Arch since I first got it. I keep my Arch up-to-date by running a full update roughly every four weeks. What is the best way to set up the new one to mirror the existing one? I'd like to have all the same packages installed, with basically the same setup, except for obvious points like the new PC having a different IP address and name on my network. My initial idea was something like the following: 1) Install a basic Arch system on the new computer 2) Get the list of installed packages from the old computer (I think there is a way to do this using Pacman?) 3) Somehow get Pacman to install the same list of packages on the new computer. 4) Tidy up any config files that need changing - e.g. for the aforementioned IP address / computer name stuff. 5) Copy over my /home directory 6) Copy over various data directories/partitions I have -e.g. I have a whole bunch of images stored in a separate partition that I'll want to copy over. The stuff in here is all self-contained so should be easy enough to copy over. I found thread https://bbs.archlinux.org/viewtopic.php?id=171726 which had some ideas as well - the above is sort-of based on stuff I found on there along with my own thoughts. Does that look like a reasonable plan? Anything I might be missing? Thanks for your attention. Regards, Spencer

7 9

libgda 6.x PKGBUILD suggestion
by Rafael Fontenelle 01 Nov '22

01 Nov '22

Hello 'community' package maintainers, libgda 5.2.10 is in the 'community' repository, and the latest version is 6.0.0. Its build system was changed to Meson along with some new features added and other deprecated disabled, besides being a soname bump. So, I understand how demotivational it is for the package maintainer. Personally, I'm a GNOME's Gtranslator user and it can not be updated to its 42.0 release without libgda6. I tried patching gtranslator's source to use libgda5 but then I faced a libsoup2-libsoup3 conflict due to its dependencies ("libsoup3 symbols detected. Using libsoup2 and libsoup3 in the same process is not supported."). Because of that, libgda6 is the way to go. So this email is to suggest a 'libgda6' PKGBUILD that was prepared and made available in the unofficial repository FCGU — notice the different pkgname to avoid replacing libgda which could cause dependency breaks. libgda6 PKGBUILD: https://codeberg.org/fabiscafe/fcgu/src/commit/ec5c32aab695886a5d373631ec01… Please kindly consider making this release available. Thanks in advance. Best regards, Rafael

1 0

Re: info: a few out of date packages
by Genes Lists 01 Nov '22

01 Nov '22

On 11/1/22 08:33, Jayesh Badwaik wrote: > `gnupg` 2.3 is the unstable version. 2.4 would be the stable version. > (FYI - you may want to use reply-to-list as mailman now preserves headers for DKIM) Hi and Nope that's not correct [1] : Three different series of GnuPG are actively maintained: - Version 2.3 is the current stable version with a lot of new features compared to 2.2. This announcement is about the latest release of this series. - Version 2.2 is our LTS (long term support) version and guaranteed to be maintained at least until the end of 2024. Only a small subset of features from 2.3 has been back-ported to this series. See https://gnupg.org/download/index.html#end-of-life - Version 1.4 is only maintained to allow decryption of very old data which is, for security reasons, not anymore possible with other GnuPG versions. [1] https://lists.gnupg.org/pipermail/gnupg-announce/2022q4/000476.html

1 0

info: a few out of date packages
by Genes Lists 01 Nov '22

01 Nov '22

As always I am enormously thankful for all the efforts and time spent by the entire arch team keeping arch active and live and the best dang distro available (imho). That said, and to be helpful, periodically I create a little list of out of date packages, some more than others and many didn't make the cut, and it's entirely my view of priorities :) Thank you again. gene Repo Current Vers Latest Package Vers Date Vers Date Age Age python-pynacl 1.4.0 200525 1.5.0 220107 592 298 wget 1.21.3 220226 2.0.1 220501 64 184 dracut 56 220218 57 220619 121 135 [0] llvm 14.06 220622 15.0.3 220818 57 75 procps-ng 3.3.17 210201 4.0.1 220821 566 72 [1] ghc 9.0.2 211221 9.4.2 220822 244 71 [2] shadow 4.11.1 220102 4.12.3 220822 232 71 doxygen 1.9.3 211231 1.9.5 220826 238 67 nmap 7.92 220807 7.93 220901 25 61 bash 5.1.016 220104 5.2 220926 265 36 linux-firmware 20220913 220913 221012 221012 29 20 dkms 3.0.7 220927 3.0.8 221025 28 7 linux-api-headers 5.18.15 220729 6.0.6 221029 92 3 [3] make 4.3 220119 4.4 221031 285 1 systemd 251.7 221024 252 221031 7 1 [4] [0] Asked for signed tag - seems currently ignored upstream https://github.com/dracutdevs/dracut/issues/1850 [1] dates approx [2] there are lots of haskell packages using this [3] should this be removed? [4] always very,very up to date, here only as it's so important. Version Age = days between arch version and latest version Latest Age = days from the latest version through today Special mention * gnupg 2.3.8 we're up to date on legacy version (2.2.40) but over a year out of date vs current 2.3. Believe we are still awaiting an update of some arch (dev) keys - Any news on this yet? * openssl 3.0.x great to see that builds are underway!! * python 3.11 was only released last week and already builds are underway - that's pretty awesome thank you.

1 0

Shadowing the boot disk on Arch system
by Spencer Collyer 30 Oct '22

30 Oct '22

Hi, This is sort-of a follow-on to my earlier message regarding how to best set up a new desktop to match the existing one I have running Arch. I'm intending that my new computer will have an M2 drive as the boot disk. As well as the OS it will also hold my /home partition. Because of a bad experience I had with a Windows laptop with an SSD drive installed, where the SSD got completely wiped back to its factory setting, my intention is to have a hard disk that acts as a 'shadow' of the boot disk. The idea is it will also be a bootable drive with exactly the same setup as the M2 drive. That way if the M2 drive dies, I can still boot using this backup drive and the system will look the same. My intention is to have a cron job that periodically (at least daily) synchronises the backup drive so it looks exactly the same as the main drive. Does that sound like a feasible plan? If it is, my question is what is the best way to initially set up the shadow drive? Should I do a full install on it, so almost treating it like I'm setting up a new computer, but making sure I have all the same packages and setup as the main drive. Or should i just use 'dd' to copy the whole main drive over to the backup drive? I'll make sure the two drives are the same size (or rather that the backup is at least as large as the main drive). Or is there another, better, way I haven't thought of or found while trawling the web? Thanks for your attention, Regards, Spencer

2 1

Moving root partition to another drive
by pete 30 Oct '22

30 Oct '22

Hi . Got a bit of a problem i need to move the root partition to another drive my own fault i did not make it big enough to start with but i cant face a complete rebuild of the system i have several 1Tb drives lurking but how do i switch things to a new drive . I remember doing that many many years ago but cant remember how now any help out there would be greatfull Thanks Pete

6 7

Proposal: Do not remove a package from the AUR until it is available in a stable repository
by Óscar García Amor 29 Oct '22

29 Oct '22

Hi all! Recently with the GNOME 43 update I have seen that certain packages are being removed from the AUR to go into the official testing repositories. The problem, in my opinion, is that packages are being removed from the AUR before they go into an official stable repository, which makes the packages "disappear" for a while. Not everyone has the testing repository enabled. In fact using AUR packages is not a condition for having testing, so what a user sees is that suddenly the package he/she was using disappears and he/she can't install it neither with his AUR helper nor with pacman. IMHO, I think it would be better to leave the packages in the AUR until they reach the stable repositories to avoid possible confusion. Greetings. -- Óscar García Amor | ogarcia at moire.org | http://ogarcia.me

4 4