Re: Guix usage and stability as daily driver
>Tue 08 Apr 2025 at 00:15, Matteo Valsasina wrote: Guix is similar to emacs in this respect: using stock one is kind of limiting. You really benefit of their usage when you customise your experience to your needs. > Am i missing some best practices? Getting familiar with packaging is a skill to include in your portfolio. Check out the doc on contributing and packaging, use profiles, and start creating packages, updating existing ones or even customising defaults. From that point, you’ll be in a good position to use your own channels, fixing issues, and keeping variants of your most usual sw. At this point, when you hit on an issue, you’ll be able to help fix it by precisely reporting the error you get, or by fixing it yourself in simpler cases. C. signature.asc Description: PGP signature
Re: How can empty generated files in the store happen? Has someone experienced this?
Hello Nicolas, As for your original question 'how can that even happen?', it's quite common for this to happen if disk is not unmounted cleanly, especially right after reconfigure. It's actually quite easy to 'reproduce' this is a VM by issuing reconfigure and then forcefully shutting it down. Since people usually reboot after reconfigure, it's a problem when the disks aren't unmounting cleanly... (ie the issue linked 77086) Nicolas Graves writes: > Possibly but I'm not 100% sure as I'm only able to diagnose it after the > fact. > > Here it happened as a side-effect of another bug (probably 76554, laptop > screen blank after kexec reboot). It hangs indefinitely so at some > point I have to force halt the computer from the power button. Maybe > that was the cause of the corruption I observe at reboot. So you are using luks, typed in your password, got a black screen so forced shutdown? I am thinking that your system might not have actually shutdowned properly, because otherwise it doesn't make sense to me your store files would get corrupted. If you actually booted after typing in your password, no store files would be modified as there is no guix command issued. And what was from previous boot was already saved. And if that's so, I don't think the bug you're linking is related. But we will probably never know for sure. > > It started to happen to me as soon as november / december. > I'm running shepherd 1.0.3 now so it doesn't seem fixed by that. In case you're running into this regularly, 1. check your disk with programs for that, to know if it's fine, 2. if disk is fine, do ensure, especially after reconfigures, your disk is synced, ie. issue `sync` command. Probably best to do it before any shutdown to be safe and to not lose any work. Specifically you would have to close programs manually, best the whole session, run sync and only then halt/reboot. Also see https://issues.guix.gnu.org/76959 for instructions from others on how to recover if you're unsure. Based on how much damage was done you might have to boot to a previous generation and remove the newer+gc, or even go to a live iso and delete more than that. Regards, Rutherther
Re: How can empty generated files in the store happen? Has someone experienced this?
On 2025-04-08 11:05, ruthert...@ditigal.xyz wrote: > Hello Nicolas, > > As for your original question 'how can that even happen?', it's quite > common for this to happen if disk is not unmounted cleanly, especially > right after reconfigure. It's actually quite easy to 'reproduce' this is > a VM by issuing reconfigure and then forcefully shutting it down. Since > people usually reboot after reconfigure, it's a problem when the > disks aren't unmounting cleanly... (ie the issue linked 77086) I thought that because we were using copy-on-write, the files only were actually saved on the filesystem when they were properly produced, and thus properly recorded in the gc. So it's not possible to have stronger guarantees on that issue? > So you are using luks, typed in your password, got a black screen so > forced shutdown? I don't even have the screen to type my password on LUKS ; it's just plain blank when I hit `reboot --kexec` and it hangs indefinitely, no message or nothing. > > I am thinking that your system might not have actually shutdowned > properly, because otherwise it doesn't make sense to me your store files > would get corrupted. When I reboot without kexec, my system usually complains about unmounted /run/user/1000, but I don't know if that might have something to do with it. > If you actually booted after typing in your password, no store files > would be modified as there is no guix command issued. And what was > from previous boot was already saved. And if that's so, I don't think > the bug you're linking is related. But we will probably never know for > sure. > In case you're running into this regularly, > 1. check your disk with programs for that, to know if it's fine, > 2. if disk is fine, do ensure, especially after reconfigures, your disk > is synced, ie. issue `sync` command. Probably best to do it before any > shutdown to be safe and to not lose any work. Specifically you would > have to close programs manually, best the whole session, run sync and > only then halt/reboot. > > Also see https://issues.guix.gnu.org/76959 for instructions from others > on how to recover if you're unsure. Based on how much damage was done > you might have to boot to a previous generation and remove the newer+gc, or > even go to a live iso and delete more than that. Thanks for the advice, I'll check that out! > > Regards, > Rutherther -- Best regards, Nicolas Graves
Re: How can empty generated files in the store happen? Has someone experienced this?
Nicolas Graves writes: > On 2025-04-08 11:05, ruthert...@ditigal.xyz wrote: > >> Hello Nicolas, >> >> As for your original question 'how can that even happen?', it's quite >> common for this to happen if disk is not unmounted cleanly, especially >> right after reconfigure. It's actually quite easy to 'reproduce' this is >> a VM by issuing reconfigure and then forcefully shutting it down. Since >> people usually reboot after reconfigure, it's a problem when the >> disks aren't unmounting cleanly... (ie the issue linked 77086) > > I thought that because we were using copy-on-write, the files only > were actually saved on the filesystem when they were properly produced, > and thus properly recorded in the gc. Yes, that is exactly what happens. But it's not relevant to this. You even have them properly recorded as you can find them with guix gc --verify. (meaning the database knows about them) But on the other hand it's not responsibility of Guix to actually make sure the files are written to the disk itself. It just makes sure what currently is on the filesystem is fine. If filesystem -> disk sync fails somewhere, that's an issue. > > So it's not possible to have stronger guarantees on that issue? No, it's never possible to have guarantees on file corruption. Ie. for disk problems or when file system is not unmounted before halting. This can cause all sorts of issues, one of them being only the inode is saved, but the contents actually not. > >> So you are using luks, typed in your password, got a black screen so >> forced shutdown? > > I don't even have the screen to type my password on LUKS ; it's just > plain blank when I hit `reboot --kexec` and it hangs indefinitely, no > message or nothing. As was said in the issue, the screen will remain blank, but you type your password anyway. You don't see the prompt, but it's there, just not visible. I actually did try this myself yesterday and it was working. > >> >> I am thinking that your system might not have actually shutdowned >> properly, because otherwise it doesn't make sense to me your store files >> would get corrupted. > > When I reboot without kexec, my system usually complains about unmounted > /run/user/1000, but I don't know if that might have something to do with > it. Yes, that could be the issue, it doesn't sound good. The unmounting should finish without errors. Regards, Rutherther
Re: Guix usage and stability as daily driver
Hi Matteo, On Tue, Apr 08 2025, Matteo Valsasina wrote: > With lot of software a problem started to show up. This is a bit of a sore point for many (and what I'm about to write is also disputed) so I'll try to keep it short: Many Guix contributors are software developers with above-average talents and skills. They find it (1) more exciting to write great software than to work on bugs. Guix is an unusual creation in that it is not just an operating system with lots of bugs (at least for me) but also a package manager with very high academic value. Guix benefits enormously from those talents and skills. Compounding the problem, there has been (2) a deluge of bugs due to Guix's rising popularity---I personally think it's the next Debian---and an ever-growing number of packages. Unfortunately, there are (3) not nearly enough contributors to deal with the resulting responsibilities, but the project is somewhat in denial about being the victim of its own success, to use a beaten phrase, and in any event no one knows what to do about it. There are also other factors such as (4) a lack of organization, (5) an unwieldly and impractical decision-making process, (6) a widespread fear to trigger rebuilds or otherwise screw up a complicated piece of software, (7) a desire keep tight control for security reasons, and (8) a lack of resources in a project driven by volunteers. It will probably take a crisis to electrify the members so they make the compromises that will be needed to take Guix to the next level. Unfortunately, it means you have to wait. Thanks for using GNU Guix! It was heart-warming to read your kind but concerned message from a fellow software connoisseur! Kind regards Felix
Re: How can empty generated files in the store happen? Has someone experienced this?
Tobias Geerinckx-Rice writes: > Anecdotally, empty store files happen a surprising lot. Surprising to me, > anyway. Not only on btrfs. Can confirm, I've had this twice in the last 3 months. And I did all the usual btrfs/smartctl checks both times, so it doesn't seem to be a filesystem issue either. (Nor do I tend to reboot right after a reconfigure, for that matter.) It's definitely one of the more annoying bugs. Not too hard to recover from, as with most issues on a Guix System, but still annoying. The kind I would definitely find unacceptable from distros with more resources, like Debian or Fedora. > It's been ages since I've used anything but Guix, but is this failure mode > really as common with other package manglers? Does anyone have anecdata to > that effect? Anecdatally, I've had something like this happen once in the two years during which I used pacman, and that turned out to be a btrfs file corruption issue that I could have caught months in advance if I'd been paying attention. I've never had it on any other package manager. So from my experience, it's far less common. But then, repairing it was MUCH more annoying, because pacman will notice these things mid-update, and then the entire rest of the update will fail catastrophically, messing up your entire system. I really appreciate Guix being transactional :) (also, package "manglers", heh heh heh)
Re: adding Samba to config.scm
gfp writes: > Hi, > > I added 3 sentences in my Samba part > to my config.scm: > > guest account = gast > create mask = 0775 > directory mask = 0755 > > (service samba-service-type (samba-configuration > (enable-smbd? #t) > (config-file (plain-file "smb.conf" "\ > [global] > map to guest = Bad User > logging = syslog@1 > > [public] > browsable = yes > path = /home/gfp/public > read only = no > guest ok = yes > guest account = gast ;hinzugefügt > create mask = 0775 ;hinzugefügt > directory mask = 0755 ;hinzugefügt > guest only = yes\n") > > > But it is impossible in Guix to add a user like this in Debian: > > sudo adduser someusername > > > Once the user exists, create a samba login: > > sudo smbpasswd -a someusername > > > So what do I have to do in Guix to add a new user and create a samba > login for that user? > Add the user to your system configuration next to your own user and reconfigure. Then run smbpasswd like you said. > thanks > > Gottfried > > > > Am 25.03.25 um 07:35 schrieb Ignas Lapėnas: > >> I'm guessing you might not have a "Guest account", to control which >> system directories it has access to. >> >> https://superuser.com/questions/1081542/how-to-allow-guest-access-in-samba >> signature.asc Description: PGP signature
Re: Guix pull with root user?
Hi Daniel, On Tue, Apr 08 2025, Daniel Hatton wrote: > I guess this advice doesn't apply to me. It does. I also use "su -" but never root's Guix, which is different from that of the other users, because I reconfigure with "guix deploy." Kind regards Felix
Re: Guix usage and stability as daily driver
Hi Cayetano On mar, apr 8, 2025 at 10:52 Cayetano Santos wrote: >>Tue 08 Apr 2025 at 00:15, Matteo Valsasina wrote: > > Guix is similar to emacs in this respect: using stock one is kind of > limiting. You really benefit of their usage when you customise your > experience to your needs. > Emacs is what brought me here in the first place I can see the similarities >> Am i missing some best practices? > > Getting familiar with packaging is a skill to include in your portfolio. > Check out the doc on contributing and packaging, use profiles, and start > creating packages, updating existing ones or even customising defaults. > From that point, you’ll be in a good position to use your own channels, > fixing issues, and keeping variants of your most usual sw. > That's on my todo list, but can't see it in the near time. Still lot to study to understand packaging Thanks for the advice Matteo > At this point, when you hit on an issue, you’ll be able to help fix it > by precisely reporting the error you get, or by fixing it yourself in > simpler cases. > > C.
Re: Idiomatic way to run Emacs 30.1?
Emacs 30.1 is available on "emacs-team" branch, so, the best way is probably to wait until it's merged. You could also use package inferiors to add it, but I'm not sure if that is a good idea.
Re: Idiomatic way to run Emacs 30.1?
Hi, On April 8, 2025 10:47:15 PM GMT+02:00, Snikta wrote: >Hi, > >I want to run the latest stable version of Emacs, but I'm not sure how to >achieve this on Guix. > >I've tried: guix shell emacs ---with-version=emacs=30.1 > >But that doesn't work. > This is just a method to replace the source, it wont work with more complex changes, those might get introduced in major versions. >Should I create a new package that inherits from emacs and then change version? That would be a possibility if no one has packaged it yet, and it would require a complex understanding of what changed between the versions in the build steps. But it is already packaged, so there is no need to do that. To run it in a shell, you can do `guix time-machine --branch=emacs-team -- shell emacs` and possibly add more emacs packages to the shell here. > >Best regards, >Snikta > Regards, Rutherther
Re: adding Samba to config.scm
Hi, 1. it is impossible to change the password in my Samba new user account: gast I even tried it in the new user account: gast In MATE it said: could not find file In KDE it said nothing, no error shown. So what is exactly the wording? sudo smbpasswd -a gast than enter the new password? 2. now I have a samba user: gast I can log in (with the initial password "firsttimepass") but it didn´t help me to enter my public folder, created in my user account, in opening my VM. error message: no shared folders found So until now no success What else do I have to do? thanks Gottfried Am 08.04.25 um 07:07 schrieb Ignas Lapėnas: now I have a second account with a password. I don´t know this password to log into it. (password (crypt "firsttimepass" "$6$abc")) how can I find out the password? The "firsttimepass" will be the "initial" password of the new account. After that I just usually use passwd to change the account password into something not written in cleartext. The "guix system reconfigure" command does not change the password for existing users, only sets them for the created ones. OpenPGP_0xD9E413C6C4BB32CE.asc Description: OpenPGP public key OpenPGP_signature.asc Description: OpenPGP digital signature
Re: XFCE Thunar application launch error
Hello, Mihael Verček writes: > Hi! > > When I'm trying to open a program from Thunar (like opening a pdf) > nothing happens and in logs I get: libgvfscommon.so: undefined symbol: > g_once_init_enter_pointer > > Opening programs from whiskermenu or terminal works as a workaround. Seems that a program is loading so library of wrong version that is ABI incompatible. This happens in thunar, because thunar is wrapped with GIO_EXTRA_MODULES variable. To fix it, you can sync your system+home+guix package versions (just upgrade them with the same guix version). Then all your programs should be using the same libgvfscommon.so and the problem will be gone. There is a GCD that tackles this issue of environment variables leaking, but currently there is no good workaround available. > > Here are some links from the net about the issue: > - https://www.mail-archive.com/bug-guix@gnu.org/msg40898.html > - > https://discussion.fedoraproject.org/t/importerror-lib64-libdnf-so-2-undefined-symbol-g-once-init-enter-pointer/104968 > > How should I go about this? > > Best regards > Mihael
Re: Guix pull with root user?
Hi Daniel, Daniel Hatton writes: > On 23/03/2025 15:50, Rutherther wrote: > >> Yes, exactly. sudo will use the guix from your user's path, so root's >> guix is not used for this. > > Hmm. I usually use "su -" rather than "sudo" when I need root > privileges (mainly because I have a history of creating overcomplicated > /etc/sudoers files such that I can never predict whether I'm allowed to > do any given task with sudo). Hence, I guess this advice doesn't apply > to me. Am I alone in this? With su -, you won't have user's guix in path, but that doesn't have to stop you from using it. You can either try making some sort of a utility that will copy the PATH (like sudo does), or just refer to it with full path, ie `/home/user/.config/guix/current/bin/guix`. So you still of course don't have to use root's guix. But you're right that my original suggestion to just run guix normally (sudo guix...) doesn't hold here. Another possibility would be to just use plain su without login shell, but that way you might end up with root owned cache files in your user's profile, so you decide. Regards, Rutherther
Re: How can empty generated files in the store happen? Has someone experienced this?
Hiza, On 8 April 2025 16:48:37 UTC, Rutherther wrote: >But on the other hand it's not responsibility of Guix to actually make >sure the files are written to the disk itself. It just makes sure what >currently is on the filesystem is fine. Anecdotally, empty store files happen a surprising lot. Surprising to me, anyway. Not only on btrfs. It's been ages since I've used anything but Guix, but is this failure mode really as common with other package manglers? Does anyone have anecdata to that effect? (Then again, I can't remember this ever happening to me, and that's on bloody bcachefs, which corrupted my store database monthly for about the first year.) >If filesystem -> disk sync fails >somewhere, that's an issue. Absolutely. Kind regards, T G-R Sent on the go. Excuse or enjoy my brevity.
XFCE Thunar application launch error
Hi! When I'm trying to open a program from Thunar (like opening a pdf) nothing happens and in logs I get: libgvfscommon.so: undefined symbol: g_once_init_enter_pointer Opening programs from whiskermenu or terminal works as a workaround. Here are some links from the net about the issue: - https://www.mail-archive.com/bug-guix@gnu.org/msg40898.html - https://discussion.fedoraproject.org/t/importerror-lib64-libdnf-so-2-undefined-symbol-g-once-init-enter-pointer/104968 How should I go about this? Best regards Mihael
Re: Guix usage and stability as daily driver
Matteo Valsasina writes: > With lot of software a problem started to show up. > System is less "stable" (pass me the expression). > Sometimes i use software after 1 month to discover it's not working any more. Are you using a package manifest? If not, that may be an option. It used to be similar for me until I started to keep a manifest and always update with guix package --fallback -c8 -M3 -k -m ~/guix.manifest This ensures that packages are always updated together with their dependencies and it kept my system far more stable. To generate a manifest, just use guix package --export-manifest > ~/guix.manifest Best wishes, Arne -- Unpolitisch sein heißt politisch sein, ohne es zu merken. draketo.de signature.asc Description: PGP signature
Re: Guix pull with root user?
On 23/03/2025 15:50, Rutherther wrote: Yes, exactly. sudo will use the guix from your user's path, so root's guix is not used for this. Hmm. I usually use "su -" rather than "sudo" when I need root privileges (mainly because I have a history of creating overcomplicated /etc/sudoers files such that I can never predict whether I'm allowed to do any given task with sudo). Hence, I guess this advice doesn't apply to me. Am I alone in this? -- Kind regards, Dan Hatton Dr. Daniel C. Hatton E-mail: dan.hat...@btinternet.com SIP:dan.hat...@sip.linphone.org Signal: dch.28 OpenPGP_signature.asc Description: OpenPGP digital signature
Re: Guix pull with root user?
On Sun, 2025-03-23 at 10:04 -0700, Vagrant Cascadian wrote: > At some point years ago the standard advice switched from "sudo -E > guix > system reconfigure ..." to "sudo guix system reconfigure ..." but > maybe > somewhere a regression was introduced, or it was subtle enough that > nobody noticed? I only noticed recently running on some diskspace > constrained systems... I searched the history of the documentation and then the repository for mentioning of "sudo -E guix …" But it seems that the recommendation has not made it into the documentation in the first place ever since the repository started, which is some time in 2012. $ git log -p -- doc/ | grep -F "sudo -E guix" $ git log -p | grep -F "sudo -E guix" When I started using guix, I just followed the documentation. Maybe, the documentation should be updated? It is probably not only wasted space but also resources and time to create a duplicate of the checkout, which the user has. $ sudo du -sh /root/ 865M/root/ $ sudo du -sh /root/.cache/guix/checkouts/pjmkglp4t7znuugeurpurzikxq3tnlaywmisyr27shj 7apsnalwq 865M/root/.cache/guix/checkouts/pjmkglp4t7znuugeurpurzikxq3tnlaywm isyr27shj7apsnalwq $ du -sh .cache/guix/checkouts/pjmkglp4t7znuugeurpurzikxq3tnlaywmisyr27shj7apsna lwq/ 700M.cache/guix/checkouts/pjmkglp4t7znuugeurpurzikxq3tnlaywmisyr27 shj7apsnalwq/ Is this expected? Why would the same checkout be different? The size is different. The files also differ. Roman signature.asc Description: This is a digitally signed message part
Re: How can empty generated files in the store happen? Has someone experienced this?
Tobias Geerinckx-Rice writes: > Hiza, > > On 8 April 2025 16:48:37 UTC, Rutherther wrote: >>But on the other hand it's not responsibility of Guix to actually make >>sure the files are written to the disk itself. It just makes sure what >>currently is on the filesystem is fine. > > Anecdotally, empty store files happen a surprising lot. Surprising to me, > anyway. Not only on btrfs. > > It's been ages since I've used anything but Guix, but is this failure mode > really as common with other package manglers? Does anyone have anecdata to > that effect? I would expect this behavior with Guix much more commonly as it tends to create a lot of new files. > > (Then again, I can't remember this ever happening to me, and that's on bloody > bcachefs, which corrupted my store database monthly for about the first year.) > >>If filesystem -> disk sync fails >>somewhere, that's an issue. > > Absolutely. > > > > Kind regards, > > T G-R > > Sent on the go. Excuse or enjoy my brevity.
Re: How can empty generated files in the store happen? Has someone experienced this?
On 2025-04-08, Tobias Geerinckx-Rice wrote: > On 8 April 2025 16:48:37 UTC, Rutherther wrote: >>But on the other hand it's not responsibility of Guix to actually make >>sure the files are written to the disk itself. It just makes sure what >>currently is on the filesystem is fine. > > Anecdotally, empty store files happen a surprising lot. Surprising to > me, anyway. Not only on btrfs. > > It's been ages since I've used anything but Guix, but is this failure > mode really as common with other package manglers? Does anyone have > anecdata to that effect? I know in Debian dpkg makes fsync calls afer many operations... If guix does not already call fsync or related system calls ... maybe it should? If it already does, maybe there are more places where it should call fsync? There will certainly be a performance hit, as a tradeoff for increased reliability... it would not fundamentally solve the problem, but it might significantly reduce the risks. There is a library that vastly speeds up dpkg operations by essentially disabling fsync... for things like initial installs or ephemeral chroot environments, where the problems resulting from datta corruption are far less significant: https://www.flamingspork.com/projects/libeatmydata/ live well, vagrant signature.asc Description: PGP signature
Re: Guix usage and stability as daily driver
On mar, apr 8, 2025 at 04:52 "Dr. Arne Babenhauserheide" wrote: > > Are you using a package manifest? If not, that may be an option. > > It used to be similar for me until I started to keep a manifest and > always update with Hi Arne, thanks for the advice. I used a manifest until a month ago. Now all packages are declared in the home-config.scm, as my understanding should have the same advantages as a manifest Matteo > > Best wishes, > Arne
Re: adding Samba to config.scm
Hi,
> So what is exactly the wording?
sudo -E smbpasswd -a gast
and then the new password
Atleast that is what I did on VM
You also should not need to login to the guest account. Failing to
login should automatically use the gast user account as the user
connected.
Setting is controlled via `map to guest = Bad User`
> What else do I have to do?
Nothing that I can think off. You have the unix account added via
(operating system ...)
You also have the samba user added via
smbpasswd
I mean it's possible to check if the folder itself is readable/writeable
by the user, but I kinda doubt that's where the issue lies.
\\{server}\public should be the UNC
You also do now seem to hit any special sections (there are 3 global,
printers, homes)
https://www.samba.org/samba/docs/current/man-html/smb.conf.5.html
I'm pretty sure you've checked the logs. If not maybe there are some
issues that are registered, although not visible...
cd /var/log/samba
--
Hope that helps,
Ignas Lapėnas
Re: Guix usage and stability as daily driver
On mar, apr 8, 2025 at 06:41 Felix Lechner wrote: > This is a bit of a sore point for many (and what I'm about to write is > also disputed) so I'll try to keep it short: > > Many Guix contributors are software developers with above-average > talents and skills. They find it (1) more exciting to write great > software than to work on bugs. Guix is an unusual creation in that it > is not just an operating system with lots of bugs (at least for me) but > also a package manager with very high academic value. Guix benefits > enormously from those talents and skills. > Hi Felix, thanks for pointing it out. I am aware and grateful of the high technical community and volunteers. Every time i read somthing like the documentation or other articles, posts, ecc about guix i find it really precise, concise and technical or obscure (for me not knowing what they are about and this gave me the oppportunity to discover more technologies) Or sometimes both.. Best regards Matteo
Idiomatic way to run Emacs 30.1?
Hi, I want to run the latest stable version of Emacs, but I'm not sure how to achieve this on Guix. I've tried: guix shell emacs ---with-version=emacs=30.1 But that doesn't work. Should I create a new package that inherits from emacs and then change version? Best regards, Snikta OpenPGP_0x8CF8982C455FAB05.asc Description: OpenPGP public key OpenPGP_signature.asc Description: OpenPGP digital signature
