Re: I installed guix with a swap and ... my swap disappeared

[Tobias Geerinckx-Rice]

Hi Jean-Michel,

Try running

# sudo mkswap /dev/sda3

to recreate the swap area.  In fact, run it now before continuing.  See the 
last paragraph.

I think that both of your problems have the same cause, and one that's easily 
fixed.

To resume from hibernation, Linux needs to be told where the hibernation image 
was stored.  It does not do this by itself.  It also won't erase stale 
hibernation images, which is why you can't swapon.

You need to add

(kernel-arguments (list "resume=/dev/sda3"))

to your system configuration, or append that string to the list if you already 
have a kernel-arguments field.  Then reconfigure for the boot loader to be 
updated.

Passing "resume=here-the-uuid-numbers-and-letters" instead of a device file 
name *should* also work, but it's not something I heavily tested since adding 
hibernation support.  I've always used device file names.

Until you can successfully hibernate and resume, you MUST run that mkswap at 
the top of this message after each failed attempt.  You risk total corruption 
of all mounted file systems if you don't.

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: Laptop won’t boot after Guix install

[Tobias Geerinckx-Rice]

Hi Ian,

If you want BOOTX64.EFI, to work around broken/forgetful firmware NVRAM, search 
the manual for the 'grub-efi-removable-bootloader'.

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: A simple question about the Guix logo.

[Tobias Geerinckx-Rice]

(Well there is the Guix/RISC-V 'logo' that I made in twelve seconds to slap on 
the Matrix room.  But everything else is lies.)

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: A simple question about the Guix logo.

[Tobias Geerinckx-Rice]

Dear me,

On 26 December 2023 14:19:32 UTC, DieRosedesNordens via  
wrote:
>It is notorious that the logo of this operating system called Guix is a letter 
>V with a pair of horns on the sides.

You've been bamboozled.  There are no letters or other occult symbols hidden in 
the Guix logo.  Only a stylised gnu, as a nod to the GNU logo.

It was chosen because it looks awesome and super cool.

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Error when installing package, CMAKE_Fortran_COMPILER could be found

[Tobias Geerinckx-Rice]

close 67416
thanks

(The above is for the Debbugs control bot.  The bug-guix@ address is for 
reporting shortcomings in Guix.  For general support, please use help-guix@.  
I've moved this discussion there.)

Hi Pierre,

It appears that you haven't given the package a FORTRAN compiler to find.  Try

  (native-inputs (list gfortran))

to add one to the build environment.  Do not waste time 'guix install'ing 
anything, this will have 0 effect.

If the build system still won't find gfortran, try

  (arguments
`(
  ...
  #:phases
  (modify-phases %standard-phases
(add-before 'configure 'find-gfortran
  (lambda _
(setenv "FC" (which "gfortran"))

or so.  Untested example, assuming the compiler binary is also named 'gfortran'.

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: guix install error no space left no device bit there is space left..

[Tobias Geerinckx-Rice]

Hi Андрей,

On 2023-10-13 21:08, Андрей гранков wrote:

Re: guix install error no space left no device bit there is space left


That's not what I see: df -h says that there are 0 bytes left on /.  
cow-store can't retroactively fix that.


If you don't want to reboot, you could try

  # mount -o remount,size=90% /

I really recommend adding some swap before doing so, though.  2 GiB of 
RAM is not much for Guix, and 1 GiB is already wasted^Wused to store 
files.


What *I* would do in your situation is to save your generated config.scm 
(if any) to /mnt, reboot, immediately mount /mnt, remove at least 
/mnt/var/guix (if it exists), swapon any swap partitions you might have, 
start the cow-store on /mnt, then run guix system init 
/mnt/etc/config.scm.  Something like that.  I'd not spend time 
recovering a live system that's already using over 50% of RAM.



Im trying to install it to t60 using i686 iso, I used graphical
   installer then switched to tty3 and ran all the comands from last
   manual installing page


The installer should have started cow-store for you during the final 
INSTALL-SYSTEM phase.


Does this mean you ran out of space even before completing the automated 
installation?  Why did you switch to manual mode?  I feel like I'm 
missing something here.


Kind regards,

T G-R

Sent from a Web browser.  Excuse or enjoy my brevity.

Re: Development shell for diffutils does not appear to work - what am I doing wrong?

[Tobias Geerinckx-Rice]

On 2023-09-27 14:43, Tobias Geerinckx-Rice wrote:

  - *All* native package inputs, if any (in diffutils' case: perl).


s/native //, but I wanted to emphasise native, since they seldom show up 
as references in the output and are hence often build-time-only.


Kind regards,

T G-R

Sent from a Web browser.  Excuse or enjoy my brevity.

Re: Development shell for diffutils does not appear to work - what am I doing wrong?

[Tobias Geerinckx-Rice]

Hi (again) Dan,

On 2023-09-27 12:58, Daniel Littlewood wrote:

--development argument. The docs say
Cause guix shell to include in the environment the dependencies of the 
following package rather than the package itself.

and I took "dependencies" to mean "build dependencies".


Correct.


But perhaps it just means the runtime ones?


I'm not aware of an option for this even if you wanted to.


So as you say, you could download the release tarball of diffutils and
be able to run it, but not necessarily to build it from the source.


The release tarball *is* the source — of the ‘diffutils’ package.  The 
only additional native (‘build’) input needed to build it is ‘perl’, 
which is also provided by ‘guix shell -D diffutils’.



I wonder why the flag is called "development" when it doesn't allow
you to develop the package.


It's true that it doesn't provide a full-featured ‘development 
environment’, including someone's favourite editor, etc.  However, I 
don't think it's poorly named.  Language is just hard.  Calling it 
‘--build’ would confuse people who think Guix has a notion of ‘build 
dependencies’, of which there are already too many.


The point of ‘guix shell -D’ is to provide an environment where you can 
build a package, as packaged in Guix, from source, as packaged in Guix.  
Hence:


  - *All* build system (‘implicit’) inputs such as a compiler/toolchain, 
GNU coreutils, etc.

  - *All* native package inputs, if any (in diffutils' case: perl).

This is all you need to build diffutils as packaged: 
https://git.savannah.gnu.org/cgit/guix.git/tree/gnu/packages/base.scm#n292


If you BYO source, different from that used by Guix, you'll have to BYO 
extra development inputs it requires.


Alternatively, you could rewrite Guix's diffutils package to build from 
Git.  You'd then add extra packages like autoconf, automake, etc., to 
native-inputs, which would then be picked up by ‘guix package -D 
diffutils’.



Perhaps it refers to developing guix?


When used in ‘guix shell -D guix’, maybe… :-)


Perhaps there is a similar expression for importing everything in a
specific build system?


That's what ‘-D’ does.  It's the assumption that autotools are part of 
any build-system that's muddying the waters here.  AFAIK they are not.


Grep Guix: if a package bootstraps ./configure and friends from scratch, 
it is expected to declare autoconf & automake (and probably more) as 
inputs.  Diffutils doesn't.  If you do, that's fine, but then you're 
operating outside of the Guix diffutils package and need to add your own 
extra inputs on the command line.



Since I was mixed up between building and installing, I also wonder
whether this is what `guix build` is for. Or, equally, if I copy the
definition of diffutils in base.scm and then do `guix shell -f
my-diffutils.scm` maybe it will build itself? It certainly seems to
have all the necessary information to do so.


That's identical to ‘guix build diffutils’, no?

Kind regards,

T G-R

Sent from a Web browser.  Excuse or enjoy my brevity.

Re: Two linux users, each has different packages available

[Tobias Geerinckx-Rice]

Hi Richmond,

Did the first user 'guix pull' recently?  That is also per-user.

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Expect bayfront (bordeaux) downtime on 26 May

[Tobias Geerinckx-Rice]

Bonjour!

This is a service announcement for users of the bayfront (or bordeaux) 
substitute server, which also handles IRC logs IIRC, and probably some 
other rando services.  That's most Guix users.


The original French message follows.  The TF;DR is that networking will 
probably be severed during the 60-90 minutes starting 26 May at midnight 
CEST.  Then, everything will be perfect again.


Kind regards,

T G-R

Sent from a Web browser.  Excuse or enjoy my brevity.

 Original Message 
Subject: [hebergement-bdx-cogent] Opération programmée 00:00 CEST 
(UTC+2) 26-MAY-2023


Bonjour,

Cogent nous annonce une opération de mise à jour de leur équipements en 
tête de réseau ce qui engendrera une coupure le

00:00 CEST (UTC+2) 26-MAY-2023 de 60 à 90 minutes.

Tous nous services hébergés au datacenter seront perturbés durant 
l'intervention.


Sacha.

Re: Too many levels of symbolic links

[Tobias Geerinckx-Rice]

Roman,

Nice.

> I still don't know which symlink was causing this issue, though.

Don't take unix error messages seriously, they probably weren't intended that 
way.

The #define for the error string you saw is ELOOP, so that should answer your 
question.

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: Too many levels of symbolic links

[Tobias Geerinckx-Rice]

Hi Roman,

Roman Scherer 写道：

  /gnu/store/097dmm40lhcf777acqh5i660j4i09k85-tzdata-2022a/share/zoneinfo/posix

links to itself.


FYI, it should link to its parent:

 /gnu/store/097dmm40lhcf777acqh5i660j4i09k85-tzdata-2022a/share/zoneinfo

(So, yes, you could …/posix/posix/posix/posix/… all night long.)


Since I read I should not mess with the store, any
ideas how to get rid of that file system loop in a safe way?


You could try

 # guix gc --verify=contents,repair

and Guix should detect the mismatch and refresh tzdata from a 
substitute server.


The unsupported alternative would be to bind-mount /gnu/store 
somewhere and manually relink the file.



I already tried re-installing it, but that had no effect.


It wouldn't.

Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: Guix can't find substitutes, even though guix weather claims they exist

[Tobias Geerinckx-Rice]

Hi Timo,

Timo Wilken 写道：

$ guix describe
Generation 69   Feb 26 2023 15:09:41(current)


Nice.

You should pull and try again.  There was a bug the past week that 
had exactly this effect, and simply updating Guix will likely make 
your woes go away.


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: Newbie user question: What do the columns shown by "guix package -I" mean?

[Tobias Geerinckx-Rice]

Rodrigo Morales 写道：

Thanks! That's what I was looking for.


I'm glad to hear it!

A related question: I didn't understand the part "retain 
backwards compatibility with isatty/stderr/etc."


First of all: I worded this poorly, as is my wont.  I should have 
said ‘by using’ instead of ‘with’.  We're not trying to be 
compatible with isatty or stderr…


Wat I meant was Guix could use the ‘isatty’ Guile procedure to 
show decorate the columns only when the output is a TTY 
(‘interactive’), or print a header line to ‘stderr’ so it won't be 
captured by a pipe.


I think that any such solution would be fragile, do the wrong 
thing in at least one case, and add too much complexity just to 
sometimes add a header.


Why not simply always add a header?

Someone might be using ‘guix package -I’ in a script, and parsing 
its output.  In fact I guarantee you that several people are doing 
so behind closed doors right now!  Adding a header would break 
such scripts.


I, personally, would welcome breaking their scripts—in exchange 
for adding a ‘--format=’ option like 
we already do for ‘guix search’.  A real API that's actually 
reliable and supportable.  I think that's fair.


I just think I'd be outvoted and don't feel like starting a fight 
:-)


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: Newbie user question: What do the columns shown by "guix package -I" mean?

[Tobias Geerinckx-Rice]

Hi Rodrigo,

Rodrigo Morales 写道：
Inexperienced GUIX user here. I have a question. What's the 
formal names
for the information presented in the columns shown by `guix 
package -I'?

(see example of output below)


I answered this question on IRC as well, so I'll just be lazy & 
paste my answer here:


 rdrg109_: That's a good question! The columns are 
 documented in the manual (‘info guix’, ‘Invoking guix package’ — 
 or 
 ) 
 under ‘--list-installed’. But it's not documented by the command 
 itself. Doing so now is probably not a good idea, even if we try 
 to be clever and retain backwards compatibility with 
 isatty/stderr/etc.

 — https://logs.guix.gnu.org/guix/2023-02-26.log#205255

Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: Coreboot,T440P & GUIX installer

[Tobias Geerinckx-Rice]

Hi Jerome,

Many people run Guix [sic] System on Coreboot (including Libreboot ) 
machines.  I have 3.  It's probably the most common single firmware 
implementation amongst Guix users—if you count all the copy/paste/modified 
vendor ones separately ;-)

I also use an Atheros card (ath9k, can't say which chip) without issue.  It's a 
fact that some routers refuse to reliably connect to older/less common 
hardware.  I'm lucky not to've encountered many.

The simple connectivity test performed by the Guix installer is at 
.
  Only one of these two servers need respond.

> (git repos for GUIX ?)

No, pre-built artefacts that are opportunistically 'substituted' for the result 
of local builds, whenever they are available.  Guix can and will always fall 
back to local builds without user intervention.  Of course, these will take 
very long and require (source code) downloads of their own.

So, questions:

- Can you not continue after this red screen about substitute availability?  I 
can't test it now, but IMO this should not be fatal.  If you can't, see the 
link above for a work-around to be performed *before* this check:

  # touch /tmp/installer-assume-online

- Can you reach either server from the command line?

- You specifically mention 'resolving GitHub'.  Why the focus on address 
resolution?  You don't mention it being relevant before that part.  Can you 
download content?

- Could you try the 'latest' (daily) image, even if I'm not aware of any 
pertinent bug fixes since 1.4?  .

If the installer has emacs as I think it does, you can use eww to test Web 
browsing.  If it doesn't have wget or curl  (I really don't remember), you can 
use 'guix download URL' as a hacky substitute (...heh).


Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: open a package in the terminal

[Tobias Geerinckx-Rice]

Hi Gottfried,

The project (≈package) named 'musescore' distributes a binary (≈command) named 
'mscore':

  ~ λ ls $(guix build musescore)/bin
  mscore

Don't ask me why they didn't call it 'musescore', but in general package names 
do not map 1:1 to binary names.  That's why peeking in the /{s,}bin directories 
is helpful.

An obvious example is 'coreutils', which provides most of the commands you use 
daily, but none named 'coreutils'.

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: setting up a new printer

[Tobias Geerinckx-Rice]

Hi Gottfried,

Does 'lpstat -v' (in the 'cups' package) list 'HP_Officejet_J4500-series' 
(exactly, no typoes)?

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: Disabling unprivileged BPF by default in our kernels

[Tobias Geerinckx-Rice]

Thanks both!  I'll do so soon.

Kind regards,

T G-R


signature.asc
Description: PGP signature

Disabling unprivileged BPF by default in our kernels

[Tobias Geerinckx-Rice]

Ekaitz Zarraga 写道：
What does Debian's kconfig list for 
CONFIG_BPF_UNPRIV_DEFAULT_OFF?


I've always had this option set to Y in my own kernels, and it has 
never so much as inconvenienced me.  However, I'm not a BPF power 
user.


Does anyone know any serious and concrete drawbacks to setting 
this option in all Guix kernels, to increase default security & 
better align with other major distros?


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: Intel i7-1165G7 vulnerable to Spectre v2

[Tobias Geerinckx-Rice]

Christian Gelinek 写道：

Spectre v2: Vulnerable: eIBRS with unprivileged eBPF

[…]
Spectre v2: Mitigation; Enhanced IBRS, IBPB 
conditional, RSB filling, PBRSB-eIBRS SW sequence


Does

 $ echo 1 | sudo tee /proc/sys/kernel/unprivileged_bpf_disabled

change this?

What does Debian's kconfig list for CONFIG_BPF_UNPRIV_DEFAULT_OFF?

Guix has it *unset* (which means default *on*) which means that 
unprivileged_bpf_disabled is 0 (which means *enabled*) because 
Linux is a hot mess and nobody cares.


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: icecat warning

[Tobias Geerinckx-Rice]

Hi Greg!

Greg Hogan 写道：
"WARNING: IceCat 102 has not yet been released by the upstream 
IceCat
project.  This is a preview release, and does not currently meet 
the

privacy-respecting standards of the IceCat project."

Why?


This warning was added[0] and is still maintained[1] by Mark H 
Weaver [CC'd], an IceCat[2] maintainer.  We should probably 
(continue to) defer to his opinion.


Kind regards,

T G-R

[0]: 
https://git.savannah.gnu.org/cgit/guix.git/commit/?id=ad21d767df29f6808b4291a10c0b32bd34b6b5d4
[1]: 
https://git.savannah.gnu.org/cgit/guix.git/commit/?id=4e88a7bfcb70f25219d9f2887121ae196431a115

[2]: https://www.gnu.org/software/gnuzilla/


signature.asc
Description: PGP signature

Re: List of modules required by linux-modules

[Tobias Geerinckx-Rice]

Hi,

dabb...@gmail.com 写道：

I've stopped the kernel build after the config phase and I've
inspected the resulting .config file. There's no evidence of
CONFIG_SCSI_ISCI in there, not even "# CONFIG_SCSI_ISCI is not 
set".


CONFIG_SCSI_ISCI depends on CONFIG_SCSI_LOWLEVEL.  Is it set?

That's the only possibility I see, apart from the obvious 
CONFIG_SCSI.


Otherwise, share your (Guix-generated, final) .config.

It's not clear why this happens with my custom build and not 
with linux-libre...

Can you help me debug this issue?
Moreover, where is created the list of module required by 
"linux-modules" ?


I'm confused by your explicit reference to 
gnu/build/linux-modules[.scm], which provides a set of mechanisms 
to work with modules.  It *should* not set policies or ‘require’ 
any ones in particular.  Does it?


Do you mean ‘%base-initrd-modules’ in 
?


This static list can be frustrating, but we can easily remove 
elements from it or even replace it entirely:


 (operating-system
   …
   (initrd-modules
 ((@ (rnrs lists) remove) "isci" %base-initrd-modules))
   …)

What I suggest you do instead (because it's what I do to my 
kernels :-) is simply ship a ‘%xenomai-initrd-modules’ variable 
along with your ‘xenomai-linux’ package, keep them in perfect 
harmony yourself, and ask users to use both.


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: Qemu not creating virtual ports

[Tobias Geerinckx-Rice]

Hi Thomas,

Thomas Ieong 写道：
I checked that all the kernel modules related to virtio are 
loaded.


Apparently it's something specific to my configuration as I 
talked
to another guixer and it was working perfectly for him with the 
latest

guix and kernel.


It's not clear to me what refers to the guest and/or host.  Is the 
host a Guix System?  Something else?


If it is a Guix System, could you provide your system.scm and the 
output of ‘guix describe’?


If it's not, could you install Guix on the host (using 
http://guix-install.sh) if you haven't yet, then try using its 
qemu package rather than your distro's?


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: not about downloading videos in Guix

[Tobias Geerinckx-Rice]

Yep.  Their mark is simply GUIX, though.

(To forestall another common objection: trademark registrations are not 
case-sensitive, and conventionally rendered as uppercase.)

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: guix deploy: modinfo.sh doesn't exist error

[Tobias Geerinckx-Rice]

Hi Dmitri,

Dmitry Matveyev 写道：

/gnu/store/ch9q9w5zbvgq8srr6dscq6hdf03pfqb4-grub-efi-2.06…


Note the ‘-efi’.


…/lib/grub/i386-pc/modinfo.sh


Note the ‘i386-pc’, i.e., non-UEFI.

This GRUB thinks it's installing in BIOS/CSM mode, but was not 
built to support this, and fails.


Is the ‘minimal ISO image’ booted in UEFI mode, that is, does 
/sys/firmware/efi exist there?


If so, is the ESP actually mounted at /boot/efi?

Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: not about downloading videos in Guix

[Tobias Geerinckx-Rice]

Heyo,

[I've un-CC'd poor Gottfried.]

Akib Azmain Turja 写道：
Excuse me, what does that "®" (REGISTERED TRADE MARK SIGN) 
means?


Mostly that you took me too seriously.  Although registration *is* 
worth considering IMO.  I would like to discuss this in Brussels.



Who trademarked Guix?


Microsoft.


As far as I know, even GNU isn't a trademark...


Sure it is.  Why wouldn't it be?  It's even a registered one, at 
least in the US.  You can ask TESS.


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: how can I download a video in Guix?

[Tobias Geerinckx-Rice]

Hi again,

On 22 January 2023 13:00:46 UTC, Gottfried  wrote:
>Hi Guix,
>
>I just tried and it worked
>
>in the terminal:
>
>yt-dlp and the url of the video
>
>and it downloaded the video

Yes... that's... that is how it's documented to work.

I'm happy for your success, but please try to use packages in the documented 
way before engaging Guix® Support :-)



Happy downloadings,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: how can I download a video in Guix?

[Tobias Geerinckx-Rice]

Hi,

It's just:

  $ yt-dlp https://URL

where URL is copied from the address bar of your browser, or a hyperlink, or...

Are you having trouble with that?  If so, please share details.

yt-dlp will extract the video from the Web page, you do not have to extract the 
video URL yourself.

Some ('obscure') Web sites aren't supported, but the most common are.  All are 
moving targets so keep yt-dlp up to date.

The version of yt-dlp in Guix is generally recent but might suffer from not 
having the latest Python version.

One tip: install ffmpeg too.  It's optional, as it should be, but without it 
each YouTube video I've downloaded was reported as 'malformed'.

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: Corefreq: How to specify output in OS definition

[Tobias Geerinckx-Rice]

Hi Petr,

I think this package might predate k-l-m, in which case the honest answer is ‘I 
don't’ :-)  If it's not supported, however, it should be.

I'll try to remember to check how I use it when I get back to my laptop.  If I 
forget, complain.

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: Changing dns during installation

[Tobias Geerinckx-Rice]

I've not done so myself, but since you're at a terminal: is editing 
/etc/resolv.conf not enough?  Or does NM periodically overwrite it even when 
the connection's stable?

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: Trying to build lilypond but guix downloads firefox?

[Tobias Geerinckx-Rice]

I don't know the answer to the question, but I'll clear up this 
sidequest:


Csepp 写道：
substituting 
/gnu/store/n6w5qizdwx2kpas061dcn5gdgvaqlwr1-firefox-91.13.0esr.source.tar.xz...


[…]

Firefox is not even officially packaged, so that's the only 
thing I can

think of.


IceCat is built from Firefox ESR sources, using the ‘makeicecat’ 
script and other modifications.  If there are no substitutes for 
the modified source, this processing can happen locally, which 
would download the original.


See 
.


That said, icecat 91.13 is 3 months old.  Somebody's Guix needs an 
update.


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: Few questions regarding the installation

[Tobias Geerinckx-Rice]

Hi Wolf,

Wolf 写道：

But it is not on the path:

root@gnu ~# pvcreate
bash: pvcreate: command not found


Indeed, lvm2(-static) wasn't actually installed on the 
installation image!


Commit 0906ce7389c6654fa3d1b3b52c8330eff0256264 fixes that:

 root@gnu ~# which pvcreate
 /run/current-system/profile/sbin/pvcreate

On IRC I was advised to try guix install lvm2-static, which does 
work
(after sourcing the profile), however it does download the 
package

from the internet.


The installer image was generated using ‘guix system image’ with 
particular Guix version N.  Hence, all packages inside that image 
are those from Guix version N.  (In Guix, there is no ‘package 
repository’ distributed separately from the ‘package manager’.) 
These packages include the image's ‘guix’ package itself.


Hence, the ‘guix’ you can run in the image will always be older 
than the Guix that put lvm2-static in the image, and may contain 
older versions of some packages.


You'll have Guix N's lvm2-static in /gnu/store, but ‘guix install 
lvm2-static’ will install Guix N-1's version.  They might match, 
they might not.  I guess here they didn't.


   That seems pointless, since I already have 
   one
locally (as seen in the output of find above), so I would prefer 
to

use that one.


You can simply

 root@gnu ~# guix install /gnu/store/84n[…]-lvm2-static-2.03.1

That might still download $things for $reasons, but they won't 
include lvm2-static.



Another question is how to undo the effect of running herd start
cow-store /mnt. I did try the obvious (herd stop cow-store), but 
that

does not seem to do the trick. So, how can I do it?


This I don't know either.  I always just reboot.

Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: Persistent environment variable

[Tobias Geerinckx-Rice]

So what's the variable?

Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: package exists but not showing

[Tobias Geerinckx-Rice]

On 1 December 2022 12:56:50 UTC, SeerLite sed:
>Perhaps `guix install` should disallow installing guix to avoid these common 
>mistakes?

I'm still in favour of merging #58583.  WDYT?



Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: How do I install a file with custom permissions?

[Tobias Geerinckx-Rice]

Hi Timo,

Timo Wilken 写道：
I'm trying to patch the `wireguard-service-type' to accept 
pre-shared
keys and add them to the generated config. This all seems to 
work
fine, except that I can't get guix to generate a 
non-world-readable

configuration file.


Alas (for your plans), this is not possible.  Guix's store model, 
inherited from Nix, is a word-readable heap.


Dealing with secrets outside of the store is one area where Nix is 
‘ahead’ of Guix, in that they seem to have multiple solutions[0]. 
Very Nix.


Guix users currently use strategies similar to the second half of 
that table: the secret is placed outside of the store, not managed 
through Guix, and the Guix service/package is pointed to it at run 
time.  Every search result for ‘secrets’ in the Guix manual is 
part of such a primitive scheme.


This is how Wireguard is set up on berlin, the Guix build farm. 
/etc/wireguard/private.key was generated manually and Guix never 
deals with it.


If you want to add secrets to Guix services, you'll have to design 
a general mechanism for doing so first.  I don't have links handy 
but I'm sure there's prior discussion, perhaps even art, on the 
mailing lists.


Sorry,

T G-R

[0]: https://nixos.wiki/wiki/Comparison_of_secret_managing_schemes


signature.asc
Description: PGP signature

Re: build all dependents?

[Tobias Geerinckx-Rice]

This sometimes requires some manual fiddling


Thanks for highlighting that.  I've never encountered it.

This is similar to another recent bug report, where ‘guix show’ 
listed something as depending on a hidden package.  Or something 
like that.


How is this not a bug?

A package manager CLI that so routinely lists bogus ‘packages’ 
that break the very same CLI is silly.


We're so used to it happening in Guix that we don't notice, or 
tolerate it when we do, or pile on ad hoc work-arounds because we 
broke the obvious sed to begin with.


Two possible solutions:

- Hide hidden packages.  This makes the most sense to me.  You'd 
 still report accurate[0] numbers for ‘guix refresh -l’.


- Allow referring to hidden packages in some contexts, or maybe 
 all contexts, and keep listing them in ‘advanced’ tools like 
 ‘guix refresh -l’.


There are probably more.

Kind regards,

T G-R

[0]: Of course these numbers aren't accurate anyway, because 
inheritance.


signature.asc
Description: PGP signature

Re: What is the latest kernel release for Guix?

[Tobias Geerinckx-Rice]

Luis Felipe 写道：

I'd keep it if people don't complain about it :)


It's an intense and unexpected colour to fill an entire screen, 
but you're the designer.  I'm certainly not complaining.


I *love* Guix's colour scheme, for which I believe you're 
responsible(?), and that (new?) purple both matches and contrasts 
perfectly.  So me hap either way.


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: guix shell: error: mount: Invalid argument

[Tobias Geerinckx-Rice]

Tobias Geerinckx-Rice 写道：
Does this system have ‘uncommon’ kernel restrictions in place, 
like
SELinux or other hardening?  Is there anything printed in dmesg 
when

the container fails to start?


I forgot: which distribution is this?

Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: guix shell: error: mount: Invalid argument

[Tobias Geerinckx-Rice]

bbb ee 写道：
So /home/dev_1 (not $HOME; see error message) 


No, I was confused as well, sorry.

In the back of my mind was another user reporting they used some 
FUSE-based encryption of $HOME, and I thought you might be using 
that.  This tainted my thinking.


I have no other idea what could be happening.  What I ‘reproduced’ 
in my previous mail is probably not what is happening to you, even 
if the error messages are the same.


But it is definitely something related to your installation:


guix shell --container coreutils -- echo Elmo


should, of course, always work.

Does this system have ‘uncommon’ kernel restrictions in place, 
like SELinux or other hardening?  Is there anything printed in 
dmesg when the container fails to start?


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: guix shell: error: mount: Invalid argument

[Tobias Geerinckx-Rice]

So /home/dev_1 (not $HOME; see error message) is not a mount point?  OK, the 
name to me implied it was.

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: What is the latest kernel release for Guix?

[Tobias Geerinckx-Rice]

Yep, I just tried it out earlier today, after seeing DNS changes.  Nice!

(Is the background colour meant to indicate betaness, or do you intend to keep 
it?)


Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: guix shell: error: mount: Invalid argument

[Tobias Geerinckx-Rice]

Hi bbb,

bbb ee 写道：
I feel guix shell --container can only mount subdir in /tmp, and 
fail for

all other dir:


Your example for ‘all other dir’ just tests /home/dev_1 again, 
which looks extremely much like a mount point.  So does /mnt.  So 
no mystery there, but I do wonder what's *special* about these 
mountpoints compared to others.


What is mounted at /home/dev_1 and (presumably) /mnt on your 
system?  Anything mounted at /home?  Anything mounted *below* 
/mnt/*?


On my system:

--8<---cut here---start->8---
~ λ mountpoint /tmp
/tmp is a mountpoint

~ λ guix shell -C --expose=/tmp=/test coreutils -- echo Elmo world
Elmo world

~ λ guix shell -C --expose=/proc=/test coreutils -- echo Elmo 
world
guix shell: error: mount: mount "/proc" on 
"/tmp/guix-directory.zTa1qb//test": Invalid argument

--8<---cut here---end--->8---

Weird!  Clearly, /proc is imbued with magical Kernel magick and 
simply canno— what?  Oh:


--8<---cut here---start->8---
~ λ mount | grep /proc
none on /proc type proc (rw,relatime)
binfmt_misc on /proc/sys/fs/binfmt_misc type binfmt_misc 
(rw,relatime)

--8<---cut here---end--->8---

Surely not…

--8<---cut here---start->8---
~ λ sudo umount /proc/sys/fs/binfmt_misc

~ λ guix shell -C --expose=/proc=/test coreutils -- echo Elmo 
world

Elmo world
--8<---cut here---end--->8---

…huh.

So, ‘guix shell --container’ apparently takes issue with rbind'ing 
mounts, or something like that.


Maybe that's something Guix can handle by bind-mounting the entire 
subtree itself, or maybe it's a hard permissions error enforced by 
the kernel.  I don't know.


Someone with podman/Docker/similar experience might.

Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: ci.guix.gnu.org unreachable

[Tobias Geerinckx-Rice]

Hi Wojtek,

Wojtek Kosior via 写道：

Today various Guix commands started showing me

substitute: guix substitute: warning: ci.guix.gnu.org: 
connection failed: No route to host


Indeed ci.guix.gnu.org's IP, 141.80.181.40, is not 
accessible. Also

when trying to connect through Tor.

Anybody knows what might be wrong?


Berlin, the server hosting guix.gnu.org, issues.guix.gnu.org, and 
ci.guix.gnu.org, was down for planned maintenance (not by us).


Some nice steps were taken today to keep guix.gnu.org up despite 
the downtime, for the first time ever.


Next time we can add a status update to the home page. 
Eventually, maybe even add some kind of server status report 
facility to the ‘guix command’:


 substitute: guix substitute: ci.guix.gnu.org reports: Down for 
 scheduled maintenance.


Or maybe not.

Anyway, this maintenance window has successfully been closed, and 
ci.guix is serving tasty toots once again.


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: (length all-python-packages)

[Tobias Geerinckx-Rice]

jgart 写道：

I imagine this does not include python packages like jrnl, khal,
clikan, t-todo-manager, gajim etc?


Yep.

If you need more reliable numbers, use ('s code snippet: there 
will be a handful of false negatives (e.g., packages written in 
Python that prefer to use Makefiles, or that mix phases from 
different build systems) but that number will be insignificant.


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: What is the latest kernel release for Guix?

[Tobias Geerinckx-Rice]

Hi Stephen,

Stephen Snow 写道：
 I am 
 curious 
 what

is the current Kernel level?


6.0.8.

There is no great Web alternative to ‘guix show linux-libre@6’, 
yet, but you can use  to search for 
‘^linux-libre’.


 Is it a pace with Libre-kernel (ie 
 6.0)?


Yes.  Guix follows Linux-Libre updates within the week, on 
average.  Comfortably so.


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: Compiling vterm: cc: command not found

[Tobias Geerinckx-Rice]

> (anymore?)

Guix has never provided the 'cc' alias for GCC AFAIR.

Perhaps you were previously using 'clang-toolchain', which does symlink cc to 
clang?  That's something that we, Guix, explicitly add; not upstream.

It's unfortunate that Guix promotes a non-GPL compiler over GCC this way.

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: Regex Rules

[Tobias Geerinckx-Rice]

I recommend starting with the regex(7) man page, before comparing 
any implementation to another.



But, iirc vis regex are different from Guile's
I might be wrong on that.


‘By default, Guile supports POSIX extended regular expressions.’ — 
(guile)Regexp Functions


‘[vis] uses what POSIX refers to as “Extended Regular 
Expressions”’ — vis(1)


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: Error using gcc with trivial-build-system

[Tobias Geerinckx-Rice]

Hi Emmanuel!

On 2022-11-08 11:23, Emmanuel Medernach wrote:

I have a local package with a compile script that
I need to add to our local GUIX packages.


(Nitpick: Guix isn't a acronym.  Never has been.)


I tried to use trivial-build-system with
gcc-toolchain input, all works well when I compile
it manually in a 'guix shell'


Because ‘guix shell’ sets up a proper profile; quite unlike the 
trivial-build-system.


Using trivial-build-system here *is* the bug.  It's not a starter kit 
for building better things; it's for trivial packages that don't need 
building.


Rather than waste your time fighting that fact, use the 
gnu-build-system.  It is that starter kit.


If your tarball doesn't provide some standard GNU things like 
./configure or ‘make check’, that's quite common.  You can simply 
(delete 'configure), set #:tests? #f, etc.


Happy building,

T G-R

Sent from a Web browser.  Excuse or enjoy my brevity.

Re: "guix system init" from regular system fails

[Tobias Geerinckx-Rice]

Hi Li!

l...@dotdot.li 写道：
  The only difference is that I did not run "herd start 
  cow-store /mnt".
That service is only available when using the installer.  I also 
believe

that it is not necessary because the documentation says that


Correct.  It's only for (and on) the ‘live’ installer.

  After this /mnt/gnu/store is populated and the bootloader also 
  appears

to be installed.  However, this is not a usable system.


OK…  how exactly did you try to use it?


/mnt/{bin,etc,
home,run} are all empty (except for /mnt/etc/config.scm).  Where 
did I

go wrong?


What makes you think you did?  I've never seen these directories 
be populated by guix system init.


Did you try to boot it?  If so, please share the errors you got.

Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: look up for Guix version installed

[Tobias Geerinckx-Rice]

Gottfried 写道：

I would like to have the numbers... guix 1.3.0-32


Right.  That's what ‘guix --version’ should provide, but it's 
currently broken:


 λ guix --version
 guix (GNU Guix) 0

You can plug the commit given by ‘guix describe’ (or ‘guix system 
describe’, which can differ!) into a local Guix git checkout as a 
horrible work-around:


 λ guix describe
   […]
   guix 39e00f7
 […]
   commit: 39e00f7f6a0b80e95cf16970d201c786684e076a
 λ git clone https://git.savannah.gnu.org/git/guix.git
 λ cd guix
 λ git describe 39e00f7f6a0b80e95cf16970d201c786684e076a
 v1.3.0-26871-g39e00f7f6a0 (your ‘32’ was just a bit optimistic 
 :-)


I'm so sorry, but it technically does ‘work’…

Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: export LC_ALL=C

[Tobias Geerinckx-Rice]

Tobias Geerinckx-Rice 写道：

such at /etc/profile


…such as…


The also do not propagate


They also…

Tired,

T G-R


signature.asc
Description: PGP signature

Re: export LC_ALL=C

[Tobias Geerinckx-Rice]

Hi Gottfried,

I hope this strikes the right balance between explaining 
‘environment variables’ from scratch and answering your immediate 
question.


Gottfried 写道：
1.  Will this now overwrite my variables for ever, or only for 
some time?


Environment variables are not saved.  They can be *set* by 
configuration files, such at /etc/profile, but these files are not 
updated when you type ‘export GUIX=awesome’ on the command line.


Setting LC_ALL like this affects your current shell, and it will 
be inherited by child processes (hence why the ‘guix’ child will 
speak English after setting LC_ALL=C in the parent shell), but 
they exist purely in RAM for the lifetime of each process.


The also do not propagate to ancestor or sibling processes: 
setting LC_ALL in one terminal window has no effect on any other 
windows.  Nor will setting LC_ALL in a shell affect new processes 
you launch elsewhere, such as from your desktop menu.  Only child 
processes launched in the same shell/window will inherit it.


As soon as you close that terminal, type ‘exit’ in the (guix) 
shell, or trip over your power cable, the setting is gone.



2.  How can I set it back to my original state?


Environment variables have no built-in notion of history, or 
defaults.  They are just variables, and setting them to something 
new overwrites the old value (if any).


So:

~$ echo $LC_ALL # yours will be de_DE, I presume
en_IE.utf8
~$ LC_OLD=$LC_ALL   # save the old value
~$ export LC_ALL=C  # in with the new
~$ echo $LC_ALL # do the thing
C
~$ LC_ALL=$LC_OLD   # restore the old value

But really, in practice, I'd just close the window/shell once 
done…  they are so cheap.


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: How to avoid downloading qtbase:debug

[Tobias Geerinckx-Rice]

Hi!

kiasoc5 写道：
=guix shell qtbase= downloads the entire debug output of qtbase 
even
though I didn't write down qtbase:debug. That's hundreds of 
megabytes
of bandwith I didn't want to use. How do I avoid downloading the 
debug

output?


If this is what I suspect it is, the only way would be ‘guix shell 
--no-grafts’.  This will disable any security fixes that haven't 
yet made it to master.


I don't recommend it, beyond testing it to see if grafts are 
indeed to blame, and perhaps narrowing down how.



Is this a bug?


If it is due to grafting, it's a long-time known annoyance, yes. 
I searched (twice) to find a specific open bug report to give you, 
but found nothing.


(Almost certainly a failure on my part; don't open a new one ;-)

I say annoyance because it's nothing more for most people, but if 
bandwidth is a serious issue for you the best way to see it fixed 
is to dig in.


Csepp 写道：

guix shell always downloads every output of all the packages.


All generalisations are false, as they say:

~ λ guix shell qtbase
…
19.4 MB will be downloaded
qtbase-6.3.1  18.5MiB …
하 qtbase ~ λ exit

~ λ guix shell qtbase:debug
…
321.6 MB will be downloaded
qtbase-6.3.1-debug  288.2MiB …
하 qtbase:debug ~ λ exit

Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: Python subprocess fails to execute Pip-installed module

[Tobias Geerinckx-Rice]

Hi Pradana,

Pradana Adrinusa AUMARS 写道：

dynamically linked, interpreter /lib64/ld-linux-x86-64.so.2, for


And this file doesn't exist.

Prebuilt binary blobs don't mix well with Guix for this reason: 
they hard-code file names such as this one.


One (brand-)new work-around is

 $ guix shell -CF [PACKAGE…] [-- your command line]

E.g.,

 $ guix shell -CF -- ./hello # random binary downloaded from 
 Debian

 Hello, world!

which creates a virtual, backwards-compatible directory layout 
within the container.


This is a (glorious) hack, but it's no substitute for proper Guix 
packaging!



outputs that bash cannot find this file.


That error is criminally misleading.  ‘A’ file, but not ‘this’ 
file.  :-/


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: bash scripts in Guix question - suspend ?

[Tobias Geerinckx-Rice]

Hi jordi,

>I'm puzzled, as long as   'loginctl suspend' is working in the shell

But '~/.config/i3/i3exit loginctl suspend' isn't.

Judging by your original snippet, you've mashed two commands together: you 
meant either

  ~/.config/i3/i3exit suspend

or

  loginctl suspend

But nobody here can tell you for certain how to call your 'i3exit' script 
without knowing what's in it.

Aside: you seem to be writing a wrapper script to wrap a wrapper script to call 
a wrapper for writing to /proc.  Is that necessary?  It can make things hard to 
follow.


Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: Virtualisation: Network problems

[Tobias Geerinckx-Rice]

Andreas Enge 写道：

Am Mon, Sep 26, 2022 at 06:05:31PM + schrieb Gottfried:

gfp@Tuxedo ~$ ps -ef | grep virtlog
it comes up with:
gfp   5233  5218  0 19:45 pts/000:00:00 grep 
--color=auto virtlog

(the word "virtlog" is in red colour)
I don't know if that means virtlogd is running?

[…]
So there is one line about grep, of no interest; and one line 
showing that

the thing I am interested in, in this case shepherd, is running.


It's cosmetic, but the standard trick here is to type:

 $ ps -ef | grep '[v]irtlog'

(or any other letter — 'virtlo[g]' also works if you remembered 
too late :-)


That pattern won't match itself, so grep won't show itself.

Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: Tried to send a patch

[Tobias Geerinckx-Rice]

Adam,

Adam Kandur 写道：

I attached my previous message as a file. Thanks for your help.


Thanks, and please do try sending it to the list again.  It might 
have been a fluke.


пн, 5 сент. 2022 г. в 15:03, Tobias Geerinckx-Rice 
:

All I can say is it never hit Mailman.


That's actually not strictly true.  It never did so *from our 
perspective*, but it could have been silently swallowed.


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: Tried to send a patch

[Tobias Geerinckx-Rice]

Hi Adam,

Adam Kandur 写道：
Two days ago I tried to send a patch to guix-patches. But my 
message didn’t

appear in mail archive? Did I send it wrong?


Probably?  It's hard to say otherwise without more information.

The Guix project has no direct control over the mail server or 
access to its logs.  All I can say is it never hit Mailman.


Assuming you used this same address, what was the subject?  Do you 
still have a copy of the message?  I could ask the GNU sysadmins 
to see if they can find traces of it.


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: Store path in package definition

[Tobias Geerinckx-Rice]

Hi,

On 26 August 2022 20:39:55 UTC, Nathan Dehnel  wrote:
>Would something like this work?
>
>(string-append (assoc-ref %outputs "out") "/passff.py")

Consider this deprecated.  It's unreliable and even where it appears to work 
can break for what seems to be no reason when, e.g., cross-compiling.

Modulo typos, and indentation being weird because 'phone,

(arguments
 (list
  #:make-flags
  #~(list (string-append "HELLO=" #$output "/bin/hello"))
  #:phases
  #~(modify-phases %standard-phases
 (add-after 'unpack 'option-one
  (lambda _
   (substitute* "foo.c"
(("hello")
 (string-append #$output "/bin/hello")
 (add-after 'unpack 'option-two
  (lambda* (#:key outputs #:allow-other-keys
   (let ((out (assoc-ref outputs "out")))
(substitute* "foo.c"
 (("hello")
 (string-append out "/bin/hello"))

From what you describe it sounds like you're using a phase, but I included 
make-flags to show that they work the same way.

I used to prefer option one, but actually prefer option two nowadays, because 
it's less magical.

>https://guix.gnu.org/cookbook/en/html_node/Extended-example.html#Build-system-arguments

This should be updated.





Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: guix fails to remove the rust

[Tobias Geerinckx-Rice]

>something with ca-certificate?

You read between the lines a little too literally :-)

>guix remove: warning: at least 0.0 MB needed but only 0.0 MB available in 
>/gnu/store
>building CA certificate bundle...
>|note: build failure may have been caused by lack of free disk space

You need to GC, or otherwise make space on the file system holding /gnu.

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: Substitute not being used

[Tobias Geerinckx-Rice]

Csepp 写道：
An explanation I've heard is that the bordeaux signing key is 
not

authorized by default or something.


I'm not sure if you meant ‘that fact on its own is useless’ or ‘I 
know how to fix that, but it didn't help’.  If it was the former:


If ‘grep 7D602902D3A2DBB83F8 /etc/guix/acl’ shows output, it is 
authorised.


If it is not authorised, you can download [0] and run
‘sudo guix archive --authorize < bordeaux.guix.gnu.org.pub’.

Kind regards,

T G-R

[0]: 
https://git.savannah.gnu.org/cgit/guix.git/plain/etc/substitutes/bordeaux.guix.gnu.org.pub


signature.asc
Description: PGP signature

Re: Gnome Tracker (tracker-miner-fs-3) behaving badly

[Tobias Geerinckx-Rice]

Tracker logging every file it sees wouldn't be a good idea: apart from the 
wasted space, you'd be adding even more I/O load to the mix...

Try stracing (guix install strace) tracker's PID(s) instead.  It will generate 
more output than you need, but file names should be obvious, and you can use 
grep to further filter the output.

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: More HP AIO scanning woes [was: I express my gratitude]

[Tobias Geerinckx-Rice]

Hi Gottfried!

Gottfried 写道：
"hplip" doesn't appear anywhere, so I can't open it and to 
choose a scanner.


I probably can't fix your problem, but note that packages don't 
necessarily install a command named after them, or an ‘icon’ (as 
you mentioned in a previous mail).  Only a tiny fraction of all 
packages do both.


There is no ‘hplip’ command or programme; don't wast time looking 
for one.


The hplip package does provide several small tools, none of them 
named hplip, of which two look relevant:


- hp-scan, which throws an error message here because I don't have 
 a supported device;
- hp-uiscan, which also throws an error, but it's ~graphical~ (so 
 it might be a better tool for you—if it works!)


There are many other hp-* commands that you can see by typing 
‘hp-’ at a shell prompt.


However.

Expect a number of these tools not to work properly.  For example, 
they might look for files in the wrong place or try to write to 
the store, simply because they've never been tested.  It is quite 
possible that you are the first to ever try using them with the 
intended hardware.  You'll have to see, and report any errors you 
get, and *perhaps* we can help fix them.


The joys of using a tiny distribution: sometimes you really are 
the first person to try something :-)



Probably no Guixer has a HP printer/scanner all in one,
so nobody was able to help me.


Very possible.  I don't.

Or any other problem that hplip doesn't appear in my MATE 
Desktop.


See above; I think you're expecting something to happen that 
won't.  We'd need to add .desktop files for that to happen, but 
that would require actually testing the tools to make sure they 
are useful (IMO).


Maybe we should have ‘I own XXX hardware’ teams…

Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: guix pull: error: commit 3946540 not signed by an authorized key: 2841 9AC6 5038 7440 C7E9 2FFA 2208 D209 58C1 DEB0

[Tobias Geerinckx-Rice]

On 2022-08-06 23:06, Tobias Geerinckx-Rice wrote:

I'm not aware of a technical reason to switch branch names.


OK, having read up on the mood in #guix I think I understand the 
reference.


But I don't think that would have been a good idea, and I wouldn't have 
supported it.


Kind regards,

T G-R

Sent from a Web browser.  Excuse or enjoy my brevity.

Re: guix pull: error: commit 3946540 not signed by an authorized key: 2841 9AC6 5038 7440 C7E9 2FFA 2208 D209 58C1 DEB0

[Tobias Geerinckx-Rice]

(Aside: the immediate issue has been resolved!  Pull to your heart's 
content.  Pull for a friend.)


On 2022-08-06 22:43, Vagrant Cascadian wrote:

What a great opportunity to switch to using "main" instead of "master"
anyways. :)

I thought git was going to at some point make "main" the default new
branch name instead of "master" anyways. May as well get ahead of the
game, now that there's a technical reason to consider switching branch
names anyways.


I'm not aware of a technical reason to switch branch names.

But: Guix should not assume 'master' anywhere, and maybe we can remove 
it from %default-guix-channel too.  Logically, if "master" no longer 
occurs in the code, we should be good to go... no?  All we have to do is 
make sure the cache follows suit or is deleted.


Some effort was made (I think by civodul) to simply track the upstream 
HEAD, whatever branch that happens to be, although I don't know if that 
was ever completed.  It's possible that some assumptions snuck in that 
we won't notice until we switch.  But it should certainly be possible.


Kind regards,

T G-R

Sent from a Web browser.  Excuse or enjoy my brevity.

Re: guix pull: error: commit 3946540 not signed by an authorized key: 2841 9AC6 5038 7440 C7E9 2FFA 2208 D209 58C1 DEB0

[Tobias Geerinckx-Rice]

On 2022-08-06 20:48, Tobias Geerinckx-Rice wrote:

  guix pull: error: commit 39465409f0481f27d252ce25d2b02d3f5cbc6723
  not signed by an authorized key:
  2841 9AC6 5038 7440 C7E9 2FFA 2208 D209 58C1 DEB0


I tried a few other random things to wriggle out of this but I think 
we're stuck (which is, design-wise, probably a good thing).


Now I eat,

T G-R

Sent from a Web browser.  Excuse or enjoy my brevity.

guix pull: error: commit 3946540 not signed by an authorized key: 2841 9AC6 5038 7440 C7E9 2FFA 2208 D209 58C1 DEB0

[Tobias Geerinckx-Rice]

Hi all,

If you try to guix pull now, this is what you'll see:

  guix pull: error: commit 39465409f0481f27d252ce25d2b02d3f5cbc6723
  not signed by an authorized key:
  2841 9AC6 5038 7440 C7E9 2FFA 2208 D209 58C1 DEB0

There is and was no security risk.

This is Guix working as intended in the presence of a commit pushed 
earlier today.  The failing commit[0] is benign, and the committer did 
nothing wrong.


The commit is signed by a subkey of the main key that Guix expects, and 
it does not deal well with that fact.  This is something we'll have to 
discuss and probably fix, both in Guix and in the git push hook on 
Savannah[1].


I'm currently waiting to hear from the Savannah admins, who are the only 
ones who can roll back master for us.  I'm not aware of any way we could 
do this ourselves.  I'll follow up when it's done.


Until then, you can:

  1. Not pull.  If your Guix was relatively recent, you're not missing 
much if anything.


  2. If you must have the very latest (valid) commit, you can run:

 guix pull --commit=ad878a2c5e5313c534ccf2546cb8c978e5295ae1

 which will validate just fine.

  3. I do NOT recommend disabling authentication.  There is simply no 
benefit to that.


TTYL,

T G-R

[0]: 
https://git.savannah.gnu.org/cgit/guix.git/commit/?id=39465409f0481f27d252ce25d2b02d3f5cbc6723
[1]: Which has been deficient for years, which I've known about, and did 
nothing about.


Sent from a Web browser.  Excuse or enjoy my brevity.

Re: language server doesn't let me emote

[Tobias Geerinckx-Rice]

Tobias Geerinckx-Rice 写道：

OK.  Did you not see the hint?


If you did, what did you think it meant and how did you try to 
address what it said?  Would adding ‘at the same time’ help?


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: language server doesn't let me emote

[Tobias Geerinckx-Rice]

jgart 写道：

I only ran this™:


OK.  Did you not see the hint?

Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: language server doesn't let me emote

[Tobias Geerinckx-Rice]

Hi jgart,

jgart 写道：
substitute: updating substitutes from 
'https://ci.guix.gnu.org'... 100.0%


Which command triggered this?


0.0 MB will be downloaded
 emacs-ht-2.3  5KiB   331KiB/s 00:00 
 [##] 100.0%
guix install: error: profile contains conflicting entries for 
emacs-ht
guix install: error:   first entry: emacs-ht@2.3 
/gnu/store/pci18w9qa6740hsaqq4cxfg793sbvn6v-emacs-ht-2.3

guix install: error:... propagated from emacs-emojify@1.2
guix install: error:   second entry: emacs-ht@2.3 
/gnu/store/7y2q2aiaahkhrc7m8ykn3i7s95ypqmlh-emacs-ht-2.3
guix install: error:... propagated from 
emacs-lsp-mode@8.0.1-0.4aafe25
hint: Try upgrading both `emacs-emojify' and `emacs-lsp-mode', 
or remove one of

them from the profile.


Did you try upgrading both packages at the same time?  (guix 
package -u emacs-emojify emacs-lsp-mode, for example?)


This hint isn't bad or unclear per se, but we should probably make 
it ridiculously explicit.


Or am I conflating an obscene amount of things I don't fully 
understand here?


We're blobs of fat in a skull trying to computer.  I'd say we're 
all coping remarkably.


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: Guix as a native-inputs ?

[Tobias Geerinckx-Rice]

Hi Olivier,

Olivier Dion via 写道：

I would like to use guix as a native-inputs for a package.


As ( implies, it's a strange thing to do.  There might be a better 
solution to whatever your unspecified problem is.



Is there any issue with that?


Only that the ‘guix’ package in guix is older than the guix that 
contains it (the Guix which is building your package).  If you run 
‘(file-append #$guix "/bin/guix") shell FOO’ in the build 
environment, it might provide an older version of FOO than you'd 
get if you specified FOO as an input directly.


If you're using it in a test suite, that might not be a problem.

Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: Guix "issues" bug tracker viewer site down

[Tobias Geerinckx-Rice]

Tobias Geerinckx-Rice 写道：
Thanks!  Sadly, it's been like that for about a week now.  You 
can use

https://bugs.gnu.org/ until it returns.


[…]

The bad news is that we have to wait for a physical human to 
plug in a
physical cable, and no such humans are physically available 
(summer

etc.).


Thanks to Ricardo & Maxim, both issues have now been resolved.  We 
can now remotely recover from most(?) serious faults.


The Shepherd SSH exhaustion issue remains to be debugged.

Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: Guix "issues" bug tracker viewer site down

[Tobias Geerinckx-Rice]

Hi Andy,

Andy Tai 写道：

502 Bad Gateway


Thanks!  Sadly, it's been like that for about a week now.  You can 
use https://bugs.gnu.org/ until it returns.


We are unable to SSH into berlin to fix this.  Luckily, the other 
vital berlin services (Web site, CI, …) are still up and running.


The bad news is that we have to wait for a physical human to plug 
in a physical cable, and no such humans are physically available 
(summer etc.).


The good news is that this cable will connect to a management 
network that will allow us to remotely recover from machine 
lock-ups and lock-outs in future, without needing trips down to 
the data centre.


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: Old packages in GUIX under Debian stable

[Tobias Geerinckx-Rice]

Hi Alexander,

>$ type guix
>guix is hashed ($HOME/.guix-profile/bin/guix)

You ran 'guix install guix' or the equivalent.  The solution is simple:

1. Don't :-)  Update Guix only with 'guix pull'.
2. Run 'guix remove guix; hash guix' and verify that type now points to 
~/.config/guix/current.


Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: Testing Pre-Release Graphical Installer 1.4

[Tobias Geerinckx-Rice]

Hi,

Matias Jose Seco Baccanelli 写道：

Actually i wanted to tryout the graphical installer for testing,


Thank you!


should
i simply download the "x86-64" iso of "GNU Guix System on Linux" 
at the

"download > latest" to have the most recent version of it ?


Yep!  The ‘latest’ ISO is built from current ‘master’.

Whilst there is a ‘version-1.4.0’ branch, it's quite dated and I'd 
wager that ‘master’ is much currently closer to what will 
eventually become Guix 1.4.0.


If you have Guix installed, you can easily build[0] an installer 
from any commit you like — including your own commits.


Kind regards,

T G-R

[0]: 
https://guix.gnu.org/en/manual/devel/en/html_node/Building-the-Installation-Image.html


signature.asc
Description: PGP signature

Re: "profile contains conflicting entries..." for packages with the same version

[Tobias Geerinckx-Rice]

Hi Chris,

Chris Keschnat via 写道：


guix install: error: profile contains conflicting entries for 
glib

guix install: error:   first entry: glib@2.70.2
/gnu/store/zqy7rhv26ax5ng8ws3d3vs8gr1k1k491-glib-2.70.2
guix install: error:   second entry: glib@2.70.2
/gnu/store/1gdvmmgimszagxr0l78kcx398c5g6608-glib-2.70.2
guix install: error:    ... propagated from dconf@0.40.0
guix install: error:    ... propagated from remmina@1.4.23
hint: Try upgrading both `glib' and `remmina', or remove one of 
them

from the profile.


In general I think the solution would be to use multiple 
profiles[1]

if I had need for multiple versions of the same package?


In *general*, the solution is to upgrade both in the same 
transaction:


 $ guix install remmina glib

(because installing the latest remmina is equivalent to upgrading 
it).


However, in this particular case it won't work because remmina 
does seem to propagate a different glib than the ‘glib’ package 
exported to the CLI.


I don't have time to investigate, but it seems related to the :bin 
output, because


 $ guix install remmina glib:bin

works fine.

In general I think the solution would be to use multiple 
profiles[1]


Maybe, yes.  For a meaningful answer we'll need to know why you're 
installing glib.  Development?  Installing libraries by hand isn't 
something ‘most average users regularly’ do in Guix.



But why does this happen here? The versions seem to be the same.


But the hashes aren't.  Version numbers aren't very relevant in 
Guix (their main practical use is to make the CLI feel familiar), 
only whether the package is identical or not.


I did a "guix pull" and "guix upgrade" before and all packages 
should

be up to date.


Yes, they would be.  That would often help mitigate this error, 
but sadly not in this case.


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: Setting up a geiser development environment

[Tobias Geerinckx-Rice]

Tobias Geerinckx-Rice 写道：

stuff


For whatever reason, [0] is missing from my mailbox.  My message 
was not meant to contradict or follow up on Fredrik's.


Kind regards,

T G-R

[0]: 
https://lists.gnu.org/archive/html/help-guix/2022-05/msg00216.html


signature.asc
Description: PGP signature

Re: Setting up a geiser development environment

[Tobias Geerinckx-Rice]

Hi Einar,

Disclaimer: I'm not a Geiser guru, but I don't think it relieves 
you of importing modules?


When I run M-x geiser followed by a

 (use-modules (gnu packages))

then specifications->manifest does the needful.

If that still doesn't work for you, it might be that 
GUILE_LOAD_PATH isn't automatically set up on foreign distros.  On 
my Guix System it set to 
"/run/current-system/profile/share/guile/site/3.0" (which won't 
exist for you).


Perhaps setting it to your 
"$HOME/.guix-profile/share/guile/site/3.0" will work, or perhaps 
it will break something else…


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: info about (identifier-syntax)

[Tobias Geerinckx-Rice]

Hi Adam,

Adam Kandur 写道：
Hi guix! where i can find documentation about 
(identifier-syntax)

(what it is and what it does)?


Start with ‘info guile’ if you haven't yet.  Or was that unclear?

There's also guile-user at gnu dot org, but I don't know how 
active it is.


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: Rotate GDM login screen

[Tobias Geerinckx-Rice]

That means that extra-config takes a list, not a string.  Add 
(list …) around the entire (multi-line) xorg.conf snippet string.


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: Rotate GDM login screen

[Tobias Geerinckx-Rice]

Hi 白い熊,

白い熊 写道：
file root to specify this config in an (extra-special-file 
statement,
so that it'd be picked up on login and GDM would be properly 
rotated?


*Must* it use extra-special-file?  If so, why?  You can try it 
(untested):


 (extra-special-file "/etc/X11/xorg.conf.d/01-monitor.conf"
 "Section \"Monitor\" blah blah blah")

and see if Guix's Xorg looks in /etc/X11/xorg.conf.d at all.

Instead, what I'd use is (untested):

(services […]
 (set-xorg-configuration
   (xorg-configuration
 ;; You might have a keyboard-layout
 ;; here already, for example.
 […]
 (extra-config "
Section \"Monitor\"
   Identifier \"DSI-1\"
   Option \"Rotate\" \"right\"
EndSection")))
 …)


As a side note - and might be associated with this - can a small
resolution be specified for just the GDM login screen - because 
on
this computer the screen is tiny, the resolution big and thus 
the GDM

login screen is with super tiny font everywhere.


That I don't know, but it's a similar ‘just create a file in 
/usr!’ mess, according to several pages like 
.


I don't know how to Guixify *that*.

Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: Guix Halts After "updating substitutes from …"

[Tobias Geerinckx-Rice]

Wamm K. D. 写道：
substitute: updating substitutes from 
'https://ci.guix.gnu.org'... 100.0%


I haven't let it run for longer than an hour but it just sits 
there and
doesn't progress any further. When I run ~guix upgrade~, it 
halts,

similarly, after listing which programs it's going to upgrade.

Anyone, possibly, know what this is (or, at least, how I might 
get

around it)?


Sure: 



Use ‘--substitute-urls=https://ci.guix.gnu.org’ (i.e., remove 
bordeaux from the default list) until it's resolved.


Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: Package definition - argument, fix for wrong file format before unpack

[Tobias Geerinckx-Rice]

Tobias Geerinckx-Rice 写道：

   "-" version "tar"))


".tar", of course.

Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: Package definition - argument, fix for wrong file format before unpack

[Tobias Geerinckx-Rice]

白い熊,

白い熊 写道：
However seeing it how should I change the unpack now - to 
extract only

via tar, however from the tar.gz file which isn't a gz really?


Guix does invoke only ‘tar’, but it's tar that chokes on the 
misleading file extension and throws the error you see.


The easiest way to convince tar of its errant ways is to set the 
correct extension in the store:


 (source
   (origin
 (method url-fetch)
 (uri "https://…/mell-1.0.0.tar.gz;)
 ;; Despite the above extension, this is actually an 
 uncompressed tar file. 
 (file-name (string-append name "-" version "tar"))

 (sha256 …)))

Kind regards,

T G-R


signature.asc
Description: PGP signature

Re: gnu pack kmonad

[Tobias Geerinckx-Rice]

Hi,

Wow!  This kmonad tutorial is the first time I see generic examples 'for Linux' 
that are silently Nix-specific, rather than, say, Debian.  Ehm... 
'progress'...?  :-/ 

Guix System happens to use /run/current-system too, but does not aim to be 
Nix-compatible.  The Guix System equivalent would be

  /run/current-system/profile/bin/sleep

(I don't know if you're running Guix System or whether 'pack' is just short for 
package, not 'guix pack'.)

However: if PATH is set correctly and kmonad supports it, I strongly recommend 
just writing

  sleep 1 && setxkbmap ...

instead.  Otherwise, your configuration will work only on Guix System *and* 
only if you install setxkbmap system-wide, not with 'guix install.

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: can't connect

[Tobias Geerinckx-Rice]

*thinking noises*

Hm,

Your message did seem to imply you already got the installer ISO somehow, from 
a GNU mirror if I had to guess.  Is that correct?

I usually recommend using the 'latest' installer as 1.3.0 is over a year old 
now, but the former isn't mirrored AFAIK and the latter should work -- it's 
just old.  Unless you hit a bug during the installation you can upgrade the 
system once it's installed.

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: can't connect

[Tobias Geerinckx-Rice]

Hi,

Due to increased attacks from Russian IP space 'lately', the MDC (the research 
institute hosting most Guix hardware) has blocked parts of it.  

The Guix project proper doesn't block anyone, but appreciates the MDC's efforts 
to keep our machines safe.

There's a third-party Chinese mirror that might work for you.  Try:

--substitute-urls='https://mirrors.sjtug.sjtu.edu.cn/guix'

As it's a pure mirror, to use it you SHOULD NOT add any signing keys beyond 
those already provided by Guix.

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: Installation: Load non-free wifi firmware (iwlwifi)

[Tobias Geerinckx-Rice]

There's nothing I can add to Zimoun & Maxim's excellent answers.

I completely agree with Maxim: this is a feature, a positive choice; not a bug, 
nor some fearful refusal to hear loud 'voices' that are already wildly 
overrepresented.

If the urge to promote proprietary software/vendors/misinformation/status quo 
is so strong,   *almost the entire Internet* is available to do so.  Literally. 
 No need to insist we give up our space, or slather it with false balance.  A 
lot of us would leave.

Just a note on 'moderation': as a 'moderator': I weed out spammy spams (many, 
so many, a day) and trolly trolls (once a year in a good year).  I have never 
moderated good-faith conversation beyond that - as long as it doesn't veer into 
either.

Kind regards,

T G-R

Sent on [a touchscreen].  Excuse or enjoy my [terrible editing].

Re: How can I install gnu guix to a usb flashdrive to run as a live system?

[Tobias Geerinckx-Rice]

Hi Greg,

I've seen different definitions of 'live system', so:

How does the system you want differ from a regular

  # guix system init system.scm /my/usb/drive

?

One clear way is that AFAIR there's no documented way (yet) to pass 
'--removable' to grub-install, so you'll have to manually rename  the grub.efi 
to bootx64.efi.  Annoying but doable.

By 'older and newer laptops', do you mean that you expect Guix to install and 
maintain both a UEFI and a BIOS GRUB at once?  I don't think multiple 
bootloaders are easy to configure (yet) in Guix, but this too should be 
possible to support.

There's also the 'read-only OS image' definition of live system, as created by

  # guix system image -t iso9660 system.scm

used to create the Guix System installer.  Unless you declare, e.g., a writable 
/home partition, all changes will be lost on reboot.

My guess is that's not what you want and it's not 'self-contained' in that it 
can't update itself.

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: how can I find the terminal output

[Tobias Geerinckx-Rice]

Gottfried,

script(1) is the classic way to log your shell (or any) session to a text file.

See .

There's more fancy stuff like asciinema but that seems like overkill here.

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: How to replace %output with #$output?

[Tobias Geerinckx-Rice]

Mekeor,

> #~(list "CC=gcc")

Yes of course it works like this wh—

Oh.  D'oh.  My bad!  Glad to hear you got there in the end.

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: How to replace %output with #$output?

[Tobias Geerinckx-Rice]

On 22 March 2022 07:26:38 UTC, Mekeor Melire  wrote:
> (list "CC=gcc")
>
>With these changes I get this error:
>
>   Wrong type to apply: "CC=gcc"

That can happen only if you apply the string as a procedure, i.e. forgot the 
the LIST.

>> So does hard-coding 'gcc' over using (cc-for-target).
>
>Yes, I would also prefer to use cc-for-target. But I wanted to simplify
>my code first. Because in order to use cc-for-target, I have to use
>backticks.

Why?  Backticks are just short for UNQUOTE.  I don't see why you'd need or even 
want to use that here.  If you insist, you could write:

  `(,(string-append "CC=" (cc-for-target)))

but I'd much prefer you didn't as I think it's a step back from

(list (string-append "CC=" (cc-for-target)))

>> Very strong opinions on Scheme coding style, that boy.
>
>Please don't assume gender. But yes, you can refer to me with the "he"
>pronoun.

If you believe yourself to be the son of god, you have problems I can't help 
you fix. ;-) 
Hi Mekeor,

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: How to replace %output with #$output?

[Tobias Geerinckx-Rice]

Hi Mekeor,

Remove the quote after arguments (replace it with a plain LIST call) and add #~ 
before (modify-phases ...) to define a gexp.

Untested example:

(arguments
  (list #:tests? #f ; no tests
 #:make-flags (list "CC=gcc")
 #:phases
 #~(modify-phases %standard-phases
 (delete 'configure)
 (add-after 'unpack 'patch-makefile
   (lambda _
 (substitute* "Makefile"
   (("/usr/local")
#$output)))

Also, newlines after ( make the baby Jesus cry.

So does hard-coding 'gcc' over using (cc-for-target).

Very strong opinions on Scheme coding style, that boy.

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: Does guix not include samba service?

[Tobias Geerinckx-Rice]

Hi Tumashu,

Nope, there is no samba service in Guix (yet :-).

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: glibc-utf8-locales

[Tobias Geerinckx-Rice]

Hullo Simon,

Simon Josefsson via  wrote:
>First, I wonder if this is optimal.  There must be many machines
>(servers and embedded) where having all locales installed on is
>wasteful, but where it is useful to have the C.UTF-8 and/or en_US.UTF-8
>installed, to get minimal working UTF-8 support.  Making this hard to
>achieve for users seems unhelpful to me.  I understand the motivation
>for the patch

I don't think you do, if that's what you thought it was ;-) 

>  (packages (append (map specification->package
> '("glibc-useful-swedish-utf8-locales"

You have a (probably left-over) specification->package call but no 
"glibc-useful-swedish-utf8-locales"  package in Guix for it to find.  It does 
not search the environment.

Remove the spec and refer to your custom variable directly, e.g.:

  (packages (append (list glibc-useful-swedish-utf8-locales ...) ...))


Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: Guix website and CI is unavailable in Russia

[Tobias Geerinckx-Rice]

Mykola,

I hope this nightmare for all Ukranians ends as soon as possible.

However, I thought we'd made it clear that these lists are no place for hatred.

Those unfortunate to live inside an authoritarian state deserve as much freedom 
and compassion as its other victims.  They certainly don't deserve to be 
punished for it.

By denying them the right to facts, privacy, and security you're not merely 
punishing them: you're doing your enemy's job for free.

Anyway, back at the topic...

Others in a similar country have previously set up a 'Guix mirror' using 
specialised software: 
.  I guess it's 
worth trying from within Russia?


Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: information about new packages in guix available

[Tobias Geerinckx-Rice]

On 6 March 2022 16:05:07 UTC, Gottfried  wrote:
>thank you,
>when I run "guix package --list-available" it doesn't show me the new 
>packages, which I am looking for, but the whole list of all packages.
>
>I would like to know which new packages are included into the list of 
>available packages.
>that would be great.
>How can I reach this goal?
>
>When I do a "guix search", what options I should use.
>I always have trouble with the mentioning of options.
>
>Gottfried
>
>
>Am 04.03.22 um 22:17 schrieb Ricardo Wurmus:
>> 
>> Hi Gottfried,
>> 
>>> I would like to know when new packages, (I don't mean updates to
>>> packages), in guix are available.
>> 
>> Whenever you run “guix pull” it will print information about new and
>> updated packages.
>> 
>> You should run “guix pull” somewhat regularly anyway, so that you can
>> get fixes for security vulnerabilities.
>> 
>>> It would safe time to be informed and not look sometimes at the guix
>>> packages available in https://guix.gnu.org/de/packages/
>> 
>> I suggest using “guix package --list-available” or “guix search”
>> instead.
>> 
>

Gottfried,

The question is of course 'new compared to when', but the most sensible answer 
is given by 'guix pull --news --details'.

Kind regards,

T G-R

Sent on the go.  Excuse or enjoy my brevity.

Re: Declare swap partition in config

[Tobias Geerinckx-Rice]

On 2022-02-27 0:38, Antwane Mason wrote:

TYPE="swsuspend"


You can only swapon(2) swap partitions.  Run

  $ sudo mkswap /dev/nvme0n1p4

and reboot, and it should work just fine.

swapon(8) does a lot more than merely call swapon(2), including 
overwriting any SWSUSPEND signature[0].


Guix does not do this.

It could, either at the expense of code manually kept in sync with 
util-linux, or by embedding a copy swapon(8) in the initrd, enlarging 
it.


Kind regards,

T G-R

Sent from a Web browser.  Excuse or enjoy my brevity.

[0]: 
https://git.kernel.org/pub/scm/utils/util-linux/util-linux.git/tree/sys-utils/swapon.c#n371

Re: Android Device Mounting Detection with GVFS Help

[Tobias Geerinckx-Rice]

Jaft,

Not quite equivalent to GVFS/udisks (for me that's a feature), but I'm 
very happy with jmtpfs.


Kind regards,

T G-R

Sent from a Web browser.  Excuse or enjoy my brevity.