Re: Remove Package from Gnome Meta-Package

[宋文武]

Lee Thompson  writes:

> Hi All,
>
> Is there an analogue in Guix for NixOS's `excludePackage`? Under NixOS I
> had something like the following in my system config:
>> environment.gnome.excludePackages = with pkgs.gnome; [
>> cheese
>> epiphany
>> gnome-music
>>   ];

I think it can be done with:
```
(define (exclude-some-packages lst)
  (filter (lambda (pkg)
(not (memq pkg (list
gnome-terminal
gnome-user-docs
gnome-maps
  lst))


(service gnome-desktop-service-type
  (let ((cfg (gnome-desktop-configuration)))
   (gnome-desktop-configuration
(core-services
 (exclude-some-packages (gnome-desktop-configuration-core-services cfg)))
(shell
 (exclude-some-packages (gnome-desktop-configuration-shell cfg)))
(utilities
 (exclude-some-packages (gnome-desktop-configuration-utilities cfg)))
(extra-packages
 (exclude-some-packages (gnome-desktop-configuration-utilities cfg))
```

>From the its code, we can see that gnome-desktop-configuration use 4
lists for meta packages, which can be customized.

Hope it helps.

Re: Adding a menu entry on GRUB.

[宋文武]

William  writes:

> Hello.
>
> I'm currently trying to add a menu entry on GRUB through the Guix
> system declaration to boot another GNU+Linux system but I don't have
> much clue how to do it with this setup.
>
> The other system currently has three partitions, one for "/boot",
> one for "/" and one for "/home", I tried defining a menu entry like
> this:
>
> ```
> (menu-entry
>  (label "GNU/Linux")
>  (device (uuid
>  "d757ca6e-ae43-41b7-bc73-7e80cd270e05" 'ext4)) 
>  (linux "/vmlinuz-linux")
>  (linux-arguments
>  '("root=UUID=da4c92d2-099f-4046-ab1d-1695860f33f0
>  rw net.ifnames=0 quiet")) 
>  (initrd "/intel-ucode.img /amd-ucode.img
>  /initramfs-linux.img"))
> ```
>
> As you can see, it contains two different UUIDs, this is intended, as
> one UUID is for the "/boot" partition, the other one is for the "/"
> partition, I copied these values from the grub.cfg menu entry generated
> by os-prober during the installation of this system, and they look the
> same on the menu entry defined by Guix.

This looks fine to me.

>
> However, this entry does not work, when I try to boot the system
> from the GRUB menu, it says it couldn't find the device by UUID
> "d757ca6e-ae43-41b7-bc73-7e80cd270e05", which is the "/boot" partition.

The error seems from "search --fs-uuid --set 
d757ca6e-ae43-41b7-bc73-7e80cd270e05"

You can press 'c' in the GRUB to enter its command line, and then run
'ls -l' to see if it actually find a partition with the correct UUID.
Maybe the UUID reported by os-prober is wrong...

Re: catch2-3.5.1.drv fails to build with home reconfigure, but succeeds with build

[宋文武]

Richard Sent  writes:

> So I was wrong and it turned out to be a mainline bug, but one that was
> only present on x86 32-bit. Here's the reports:
>
> https://gitlab.com/nonguix/nonguix/-/issues/307
> https://issues.guix.gnu.org/68429
> https://github.com/catchorg/Catch2/issues/2796

Hello, catch2 had been updated to 3.5.3 in commit 584e240842, which
should works as it build fine on CI for i686-linux.

Please run 'guix pull' to update and test again, thanks.

Re: How to delete per-user profiles

[宋文武]

Jorge  writes:

> Hi!  I have removed (via Gnome Settings) user `jorge-morais' from my 
> Gentoo GNU/Linux system, but its Guix profiles remain.  How do I remove 
> them, so Guix GC can clean the store?  Can I simply remove the directory 
> `/var/guix/profiles/per-user/jorge-morais' ?

Yes, that's the right and safe thing to do.

Re: Eradicating PulseAudio

[宋文武]

Robby Zambito  writes:

> Hi,
>
> Please help me eradicate PulseAudio from my system. It causes me nothing
> but pain. My most commonly run commands according to my bash history are
> pulsemixer followed by pgrep pulseaudio, which I use to debug my audio
> issues, and pkill pulseaudio followed by herd restart pipewire which I
> use to resolve my audio issues caused by pulseaudio being started by
> Shepherd.

Hello, you can disable the autospawn for PulseAudio by
put "autospawn = no" in ~/.config/pulse/client.conf.


It was not started by shepherd but autospawn by applications (eg: pulsemixer),
for reference:
https://www.freedesktop.org/wiki/Software/PulseAudio/Documentation/User/Running/

Thanks.

Re: Trying to understand why GUIX isn't reading configuration file?

[宋文武]

Patrick Pippen  writes:

> Previously what I thought was a gdm issue, upon further investigation I
> believe this to be a guix issue or my lack of understanding guix way of
> doing things.
> My current issue is after installing guix with a guided installation. I
> preceded to install bspwm as a user, what is happening is
> I created a .xsession file and chmod +x ~/.xsession with the following
> content:
>
> sxhkd -c $HOME/.config/sxhkd/sxhkdrc &
> polybar -c $HOME/.config/polybar/config &
> exec bspwm -c $HOME/.config/bspwm/bspwmrc

make sure it start with a shebang line, which should be:

#!/bin/sh
sxhkd -c $HOME/.config/sxhkd/sxhkdrc &
polybar -c $HOME/.config/polybar/config &
exec bspwm -c $HOME/.config/bspwm/bspwmrc

>
> At first I've tried: exec bspwm
> And it still doesn't work.
>
> The issue that I'm not understanding is the bspwmrc file isn't being read
> after logging in so I have to open a terminal and do this.
>
> source $HOME/.config/bspwm/bspwmrc
>
> Looking at my config.scm file, I've noticed that services-type-desktop only
> shows XFCE. How can I get bspwm added to this file?
> Or is there a way to get bspmwmrc file read automatically?

I believe 'exec bspwm' will read bspwmrc from the right location , that
file is for bspwm not for shell (to be sourced).

>
> At the login screen, although XFCE and BSPWM are both installed there is
> only a login prompt. To run XFCE I have to remove the ~/.xsession file.
> XFCE was installed during installation of GUIX, after install as regular
> user I install BSPWM.

Okay, bspwm does have a share/xsessions/bspwm.desktop file, so if you
put it in the 'packages' field of the 'operating-system', it should
be available in the gdm login manager.

To allow selected xfce to run, use 'exec "$@"' in ~/.xsession instead of
'exec bspwm'.


Hope this will be helpful!

Re: Using ibus input methods in Emacs GTK+

[宋文武]

Luis Felipe  writes:

> Hello,
>
> I upgraded my Guix system (currently on 17fadbb) and now I can type
> Japanese again (thanks, 宋文武). However, ibus input methods are
> working everywhere (GTK, GTK+, Qt) except in Emacs GTK+, which is
> unexpected because typing Japanese in Emacs used to just work before. 
> My current workaround is to start "emacs -nw" in a GNOME Console and
> then I can use ibus in it.
>
> I thought that maybe something in Emacs 28.x had changed, but trying
> on a Debian 12 machine, I can use ibus normally in Emacs GTK+ without
> configuring anything.
>
> If you're able to type Japanese (or use ibus input methods in general)
> in Emacs GTK+, did you have to do anything special?
>

Well, Emacs use XIM, so you need set 'XMODIFIERS=@im=ibus' and run
ibus-daemon with the "--xim" / "-x" argument.

Hope this helps!

Blind people unable to install Guix System

[宋文武]

Simen Endsjø  writes:

> It was brought to my attention that blind people runs into problems installing
> Guix System, while Debian and Arch have solved this better.
>
> The problem is outlined by the user k.g.roberts, ref
> :
>
>   The installer for Guix System 1.4.0 is basically the same as I remember 
> it from
>   a few years ago; boots almost immediately into an ncurses menu for 
> locale
>   selection and it takes another few screens to get to the option for a 
> shell. No
>   other assistive technologies built into the installer itself outside of 
> what’s
>   in the kernel, so a blind installation is doable over a remote 
> connection after
>   that point, but would likely be a hassle to get going without 
> additional help
>   from a sighted person.
>
> Dave Dude mention a solution, ref 
> :
>
>   They’d need to add a console screen reader, like Fenrir or ESpeakup to 
> the
>   installation image, then have something to type, on the boot command 
> line, to
>   start the reader. Debian has done this, on its installation images for 
> a long
>   time. Arch now does something like this, too. Thanks for looking into 
> this!
>
> I’m not the right person to do something like this, just thought I should 
> report
> it in case it’s easy to fix.

Hello, I CC this email to bug-guix to open an issue now.

Thank you for the report!

Re: Newbie: Run previous version of Inkscape that exists in /gnu/store

[宋文武]

Rodrigo Morales  writes:

> [...]
> | 140M/gnu/store/qp8msw5js3gdpdlq1r8nxz4frnzc39lh-inkscape-1.1.1
> | 181M/gnu/store/3a7a5i1f4c917vlnzyy9d68s55ijiq50-inkscape-1.2.1
>
> I want to run v1.1.1 because I want to do some experiments to test the
> performance of these two versions. Preferably, I want to run those two
> versions at the same time. If that's not possible, at least I want to be
> able to switch between those two versions. How to do this?

Does run them by path from command line in a terminal emulator works?

 /gnu/store/qp8msw5js3gdpdlq1r8nxz4frnzc39lh-inkscape-1.1.1/bin/inkscape
 /gnu/store/3a7a5i1f4c917vlnzyy9d68s55ijiq50-inkscape-1.2.1/bin/inkscape

Re: Message by a novice

[宋文武]

anguriamelone--- via  writes:

> Hi to everyone.
>
> I am a novice in GNU Guix, so far I used Trisquel GN-Linux 9.0.2
> (i686), but at the end of April the support for the i686 version has
> been discontinued.

Hello, while not abandoned, i686 support in Guix is not well maintained.
Which means less packages are available compare to x86_64, and require
more time build (and maybe fixing) packages yourself instead of download
prebuilt ones ('substitutes' in guix terms).


> I would like to know what are the basic command lines corresponding to
> those of Trisquel GNU-Linux below:

I'll give a quick answer, more info can be find in the guix manual and
online videos.

https://guix.gnu.org/en/manual/en/html_node/Package-Management.html#Package-Management
https://guix.gnu.org/en/videos/2020/everyday-use-of-gnu-guix-part-one/

>
> sudo apt update

guix package --upgrade

> sudo apt upgrade

guix pull

> sudo apt install (package name)

guix package --install

> sudo apt remove --purge (package name)

guix package --remove

> sudo apt search (package name)

guix search

> sudo apt autoremove
> sudo apt purge
> sudo apt autoclean
> sudo apt clean

guix package --delete-generations
guix gc

Underhood, it's very different to apt, see the manual for details.

Hope this will help!

Re: firefox window crashes when opening dialog box on xfce

[宋文武]

"jgart"  writes:

> hi,
>
> My Firefox window crashes when opening a dialog box using xfce-service-type 
> on Guix System.
>
> Any thoughts on where I can start debugging this?

Hello, I think you can launch it via a terminal (eg: xfce4-terminal) to
see the output when crash.

(My icecat is working fine when show the open file dialog).

Re: desktop services

[宋文武]

Gottfried  writes:

> Hi,
>
> 1.
> In the manual 12.9.9 there are several desktop services mentioned
>
> But cinnamon-desktop and kde plasma-desktop is not mentioned,
> even though you can download the cinnamon and plasma-desktop packages.
>
> Are the mentioned desktop services like Gnome, XFCE, MATE only examples
> and cinnamon and plasma are not mentioned?
> or is the cinnamon and plasma desktop not available/working yet,
> even though the packages can be downloaded?

Those "desktop services" provide additional functions / integrations for
desktop environments, notable are screen backlight control, screensaver,
etc.  When a desktop service is available, it means the desktop
environment is well supported / tested.  For cinnamon and plasma, while
most seperated packages are packaged, they're not well integrated /
tested, missing some functions and need some future works.

>
> 2.
> You can download the desktop environment "XFCE 4.18.2"
> Additionally there are many other XFCE packages available for download.
>
> Are those additional packages completely separate packages
> which I can additionally install?
> or are some of them included in the  main "xfce 4.18.2" package?
>
> How can I know which packages are included in the main xfce 4.18.2
> package? Which command tells me that?
> in order not to install something twice?

Run 'guix show xfce' show what included in the xfce meta package,
basically the core desktop packages.  Additional applications or plugins
can be install seperated.
>
> Or is it no problem to install something twice, Guix will tell me that
> and handle it.

Yes, there are no problem.

Hope this help!

Re: catfish

[宋文武]

Gottfried  writes:

> Hi,
>
> I deleted catfish from my default profile
> and wanted it to install in a separate profile.
> But it failed 3 times.
>
> here the message
> can somebody check what that means?

Hello, catfish had been fixed in master (with commit 1fc66b9db).
Run 'guix pull' to get it, Hope it helps!

Re: display manager doesn't read ~/.xsession when it is a symlink

[宋文武]

Rodrigo Morales  writes:

> [...]
>   I've symlinked `~/my/guix-config/xdg/xsession' to `~/.xsession' by
>   using the following instruction in
>   `~/my/guix-config/home-configuration.scm' (see appendix for my
>   complete configuration)
>
>   ,
>   | (... some omitted lines ...)
>   |(service
>   | home-files-service-type
>   | `((".xsession"
>   |,(local-file "xdg/xsession"))
>   | (... some omitted lines ...)
>   | #+END_SRC scheme

Hello, you can use (local-file "xdg/xsession" recursive? #t) to make the
store file kept the permission bits, which would lead to a executable
~/.xsession.


Hope this helps!

Re: Help-Guix Digest, Vol 88, Issue 4

[宋文武]

Gottfried  writes:

> Hi,

Hello!

> 1.
>> GUIX_PROFILE="~/guix-profiles/emacs/emacs"
>>> . "$GUIX_PROFILE"/etc/profile
>
> As unexperienced in Scheme I am asking
>  if the second line beginning with the "dot" is separated for easier
>  read, but everything is one command?

Those 2 line are bash commands,
  GUIX_PROFILE="~/guix-profiles/emacs/emacs" will set a shell variable
  GUIX_PROFILE with a value of "$HOME/guix-profiles/emacs/emacs".

  . "$GUIX_PROFILE"/etc/profile will 'source' the /etc/profile in the
  $GUIX_PROFILE directory.

You can find them in 'info bash' at: "5 Shell Variables" and "4.1 Bash Shell 
Builtins".

>
> 2.
> I have got already many packages in my profile and
> updating takes a long time.
>
> So I am thinking of splitting of some packages to create several
> profiles and AFAIU updating with
> sudo guix system reconfigure /etc/config.scm
> it will take less time because it will not automatically update all my
> profiles at once. (Is this right?)

Yes, the system profile (via config.scm) and user profile already do a
split, and you can split the user profile into multiple profiles for
more control.

See 'info guix-cookbook' for "Guix Profiles in Practice".

>
> 3.
> I was creating a profile with "Musescore"
> and a profile with "Emacs".
> I want to create still other profiles.

No problem.
>
> 4.
> Now I still don’t understand everything concerning profiles.
> If I enable all profiles at login time
> how will it work?
Enable a profile mean source ('.' command in your first question) its
profile file to add its search-paths to the current shell environment:

  PATH for making binaries from profiles's bin directory available to the
  current shell.
  MANPATH for making manpages from profile's share/man dinectory
  available to the current shell's 'man' command.
  XDG_DATA_DIRS for making desktop applicatinons available to the
  current desktop environment (launcher, etc).
  And so on...

> How can I enter the different profiles?
If you enable all at login time (via ~/.bash_profile), there is no need
to switch.  If you didn't enable a profile, you can switch it with 'guix
shell -p' or '.' its profile.


> How do I know in which profile I am and to switch to an other?
You can run 'env' find out what profiles are activated now.

>
> e.g. I have emacs installed in my main profile.
> I have got a manifest with emacs with additional emacs packages.
> When entering this separate profile with guix shell
> I get to  guix shell (env)...
> but when I entered:  "emacs"
> it opened emacs with the package "Icicles" which I don’t have in my
> emacs manifest, only in my emacs in the main profile.
> So I concluded, it is the emacs in my main profile and not the one in
> my emacs manifest profile.
> How are both connected? Are they separated?
Run 'which emacs' return the emacs executable location, it maybe the one
in your main profile or the emacs profile.  There are some environment
variables / search-paths (EMACSLOADPATH) for emacs, emacs will load its
packages from this EMACSLOADPATH.  When combine profiles via source
('.') the search-paths are merged (see 'export' in the 'profile' bash
script),  so they're connected via those search-paths.

They can also be separated if you don't combine those search-paths.
eg: via 'guix shell --pure -p ...'


> But this emacs manifest doesn’t have a init.el file etc.
Emacs's init.el are at ~/.emacs.d/init.el (or ~/.emacs), it's out of
guix's control, so shared with every emacs.

> May be profiles are not completely separated in Guix like I understood
> until now.
Well, you can combine them (by default the system profile and user
profile are combined), or use them in a separated way.


Basic are shell, environment variables, and the effects of environment
variables.  Hope this helps!

Re: Newbie question: Why my system has so many shells.scm files?

[宋文武]

Rodrigo Morales  writes:

> Table of Contents
> _
>
> 1. The context
> 2. The question
>
>
> 1 The context
> =
>
>   Newbie user here. I'm curious about how `home-bash-configuration'
>   works, so I wanted to jump to its definition. I know that one way of
>   doing this is cloning the Guix repository and executing grep to search
>   that string. In fact, I did this and found it in
>   `gnu/home/services/shells.scm'.
>
>   Using common sense, I knew that that file also existed in my system,
>   so I was wondering how I could have accomplished the same objective
>   without having to download the Git repository, just in case some day I
>   don't have internet connection
>
>   ,
>   | locate shells.scm | grep '/gnu/store'
>   `
> [...]
>
> 2 The question
> ==
>
>   How to know which of these files is currently being used by my
>   system?  I'm sure not all of them are being used. I'm somewhat
>   familiar with how generations work, but in other words, I'd like to
>   know how to determine which of these files is the one that is being
>   used by the current generation.

First to find out what files are currently going to be used the system:

  $ guix repl
  > %load-path

I got:
```
$1 = 
("/gnu/store/sg5mf194lisl59b1bm3a7x6llf3vvhlv-guix-module-union/share/guile/site/3.0"
 "/gnu/store/qbqdkziw7222077fn4z2x5n91185ivis-guile-3.0.9/share/guile/3.0" 
"/gnu/store/qbqdkziw7222077fn4z2x5n91185ivis-guile-3.0.9/share/guile/3.0" 
"/gnu/store/qbqdkziw7222077fn4z2x5n91185ivis-guile-3.0.9/share/guile/site/3.0" 
"/gnu/store/qbqdkziw7222077fn4z2x5n91185ivis-guile-3.0.9/share/guile/site" 
"/gnu/store/qbqdkziw7222077fn4z2x5n91185ivis-guile-3.0.9/share/guile" 
"/home/iyzsong/.config/guix/current/share/guile/site/3.0" 
"/run/current-system/profile/share/guile/site/3.0"
```

In the first directory, it has:
  
/gnu/store/sg5mf194lisl59b1bm3a7x6llf3vvhlv-guix-module-union/share/guile/site/3.0/gnu/home/services/shells.scm
  
These scm files are guile modules, which will be used by the time 'guix' run.


And to find out what the current generation are using, first you need to
find out what guix version(s) being used.
For system profile, /run/current-system/provenace or
/run/current-system/channels.scm has the version (guix commit) info.
For user profile, ~/.guix-profile/manifest have it in the provenance property.

Then given the commit, how to get the files?

If you 'guix pull' before as I did, it can be:
```
cd ~/.cache/guix/checkouts/pjmkglp4t7znuugeurpurzikxq3tnlaywmisyr27shj7apsnalwq
git checkout -f ${COMMIT}
```

Or 'guix pull --commit ${COMMIT}', or git clone a new guix repository...


Hope this helps!

Re: Newbie user: Feedback on custom package definition

[宋文武]

Rodrigo Morales  writes:

> [...]
>   |  #:builder
>   |  #~(begin
>   |  (use-modules (guix build utils))
>   |  (chdir (assoc-ref %build-inputs "source"))
>   |  (install-file "org-recoll.el"
>   |(string-append
>   | #$output
>   | "/.config/emacs/libs")
>   |(synopsis "My synopsis")
>   |(description "My description")
>   |(home-page "https://example.org/";)
>   |(license gpl3+)))
> [...]
>
> 3 The questions
> ===
>
>   + What changes would you do to improve the definition of the package
> that I wrote?

I would update the synopsis, description, and switch from
trivial-build-system to emacs-build-system, and send a patch to
guix-patches.


>   + Do you manage your Emacs packages with GUIX? Could you briefly
> describe your workflow or point me to references/documentation?

Yes, I just:

1. install emacs packages (emacs, emacs-use-package, emacs-magit, etc.)
by 'guix package -i' or 'guix package -m'.

2. setup them usually in ~/.emacs.init with '(use-package ...)'.

Packages installed by guix are available to emacs like package-install
thanks to 'guix-emacs-autoload-apckages'.

Re: attempt to create a profile

[宋文武]

Kyle Andrews  writes:

> In my case I am hopeful there must be some way to extend the man and info
> paths when multiple profiles are activated, and maybe that would resolve
> my particular issue. I would love a suggestion to be placed in the
> cookbook about this.

Hello, you can merge search-paths from multiple profiles with:

--8<---cut here---start->8---
# Honor system-wide environment variables
source /etc/profile

# Merge search-paths from multiple profiles, the order matters.
eval "$(guix package --search-paths \
-p $HOME/.config/guix/profiles/a \
-p $HOME/.config/guix/profiles/b \
-p $HOME/.config/guix/current \
-p $HOME/.guix-profile \
-p /run/current-system/profile)"

# Prepend setuid programs.
export PATH=/run/setuid-programs:$PATH
--8<---cut here---end--->8---

As discussed in #20255 and #61358.

Re: problems with "gedit"

[宋文武]

Gottfried  writes:

> Hi,
>
> gfp@Tuxedo ~$ guix upgrade gedit
> The following package will be upgraded:
>gedit 42.2 -> 44.1
>
> guix upgrade: error: profile contains conflicting entries for dconf
> guix upgrade: error:   first entry: dconf@0.40.0
> /gnu/store/gkr4pbmc7320zvrnl48mjacpb3km00j7-dconf-0.40.0
> guix upgrade: error:... propagated from gedit@44.1
> guix upgrade: error:   second entry: dconf@0.40.0
> /gnu/store/0fpkqzkbwl43apj6ls4sfggj17i3nvvv-dconf-0.40.0
> guix upgrade: error:... propagated from gedit@42.2
> hint: You cannot have two different versions or variants of `gedit' in
> the same profile.

Um, seems like a bug here that upgrade didn't exclude previous
propagated packages from conflicts checking, not very sure, need check
though..
>
>
> How can I solve this problem?
> Should I uninstall "gedit" and then reinstall it?
> or are there other possibilities?

I think uninstall should works, or maybe "guix upgrade gedit dconf"?

Re: New package request: aspell-dict-bn

[宋文武]

Akib Azmain Turja  writes:

> I didn't modify any Guix files.  I put that in a file (along with
> supporting forms) and passed the file to "guix package".
>
> I have the Guix repository (as left by "guix pull"), but how can I test
> my changes?

Hello, The guix manual (info guix or better viewed in emacs) has
a "Contributing" section talk about how to do this.  Basically:

1. git clone the guix repository.
2. run bootstrap, configure and make
3. modify the correct source file
4. test it with ./pre-inst-env guix
5. git commit and git format-patch
6. send the patch to guix-patches mailing list

I think the checkout from "guix pull" can be used with "git worktree"
with some care, but I usually use a seperated git clone for contributing.

Hope this helps!

Re: New package request: aspell-dict-bn

[宋文武]

Akib Azmain Turja  writes:

> I need Bangla (Bengali) dictionary for GNU Aspell.  So I searched and
> found one on GNU FTP, and I tried to pack it, here's my attempt, works
> for me:
>
> (define-public aspell-dict-bn
>   (aspell-dictionary
>"bn" "Bengali"
>#:version "0.01.1-1"
>#:sha256
>(base32 "1nc02jd67iggirwxnhdvlvaqm0xfyks35c4psszzj3dhzv29qgxh")))

Pushed, thank you!  Well, next time you should submit a git patch to
guix-patches for better workflow...

Re: GUIX behind a proxy

[宋文武]

"Diederich, Stefan"  writes:

> Dear Sir or Madame,
>
> I installed guix with the installation script provided
>
> sudo sh guix-install.sh
>
> After this I added some the https_proxy and http_proxy to the environment in 
> the file /etc/systemd/system/guix-daemon.service:
>
> # This is a "service unit file" for the systemd init system to launch
> # 'guix-daemon'.  Drop it in /etc/systemd/system or similar to have
> # 'guix-daemon' automatically started.
>
> [Unit]
> Description=Build daemon for GNU Guix
>
> [Service]
> ExecStart=/var/guix/profiles/per-user/root/current-guix/bin/guix-daemon 
> --build-users-group=guixbuild
> Environment='GUIX_LOCPATH=/var/guix/profiles/per-user/root/guix-profile/lib/locale'
>  LC_ALL=en_US.utf8
> Environment='https_proxy=https://10.10.6.11:8080'
> Environment='http_proxy=http://10.10.6.11:8080'
> Environment='HTTP_PROXY=http://10.10.6.11:8080'
> Environment='HTTPS_PROXY=https://10.10.6.11:8080'
> [...]
>
> Is there anything I might have missed? Can someone help me with configuring 
> guix behind a proxy?
>

Hello, I think the value of https_proxy should also be
'http://10.10.6.11:8080', not https.  Unless you does have the proxy
server listen https, which usually not setup that way.

Hope it help!

Re: no C_INCLUDE_PATH env variable defined : I can't compile C applications

[宋文武]

Mohamed Amine LEGHERABA  writes:

> Thanks Ricardo for your answer.
>
> Sorry if my message was not clear,
> I don't have the env variable C_INCLUDE_PATH in my guix system and I
> should have it because I have the packages "gcc-toolchain" and
> "linux-libre-headers" in my home profile.
> When I run a instance of guix shell with these 2 packages, for example
> with the command "guix shell gcc-toolchain linux-libre-headers" I do
> have this env variable defined (in fact a guix shell with only
> gcc-toolchain works also) so I don't know why I don't have this
> variable defined in my system.

Hello, I think you're missing 'home-bash-service-type' in your home
services, which should setup bash to source
'$HOME_ENVIRONMENT/setup-environment', and 'setup-environment' will
source '.guix-home/profile/etc/profile'.  If gcc-toolchain is in the
home profile, that '.guix-home/profile/etc/profile' should have
C_INCLUDE_PATH in it.

Hope this helps!

Re: Deploy to guix machine

[宋文武]

Reza Housseini  writes:

>> Try replacing "~" with (getenv "HOME")?  I don't see anything in the
>> manual about it auto-expanding tildes like Bash does.
>
> Thanks good catch, I tried it but the error message is still the same
> must be something different. I especially wonder why it was working
> before, maybe I should check the commits in deploy...

Hello, the 'identity' field in 'machine-ssh-configuration' is for the
SSH private key, not the public key.

I guess it was working due to ssh agent was picked, but not now.

Re: bug#56778: Why Emacs is echoing message for each installed Emacs package while startup

[宋文武]

Akib Azmain Turja  writes:

> From d4b449e7c129a6d44414790ad36046e4d325a55e Mon Sep 17 00:00:00 2001
> From: =?UTF-8?q?Andr=C3=A9=20A=2E=20Gomes?= 
> Date: Sat, 21 May 2022 14:42:12 +0100
> Subject: [PATCH] gnu: emacs: Suppress loading messages.

Pushed with adjusted commit message, thank you!

Re: simple config with just startx not working for me

[宋文武]

Andreas Reuleaux  writes:

> [...]
> I want to start X on my system with startx - i.e. log into my system
> just from a tty, and then start X by hand (and i3 will be my window
> manager, I just have an .xsession file in my $HOME, where I start i3)

Hello, you can use 'xorg-server-service-type' (not documented, patch
welcome!)  with 'sx' (recommended) or 'xinit':

1. Add 'xorg-server-service-type' to the system config, and reconfigure.
   After this you'll have 'X' and 'Xorg' in '/run/current-system/profile/bin'.

2. Install 'sx' or 'xinit', config them:
   - For 'sx', create the ~/.config/sx/sxrc script, which likely ends
   with 'exec dbus-run-session -- i3' .
   - For 'xinit', create the ~/.xinitrc script and a ~/.xserverrc with:
--8<---cut here---start->8---
   tty=$(tty)
   tty=${tty#/dev/tty}
   X vt${tty}
--8<---cut here---end--->8---
 This xserverrc is needed for rootless X to start with the current tty.

3. Run 'sx' or 'xinit' from tty.  I think you also need to to be in the
   'video' and 'input' group for this to work.

Hope it helps!

Re: GNU Guix 1.3.0 released

[宋文武]

Maxim Cournoyer  writes:

> We are pleased to announce the release of GNU Guix 1.3.0!
>
> This release corresponds to 8,300 commits over almost 6 months by 212
> people.  Support for the POWER9 platform is now offered as technological
> preview.  This release adds new features, refines the user experience
> and improves performance.  It also includes many new packages and
> services along bug fixes--see below for a list of changes.
>
> Read more about today’s announcement at:
>
>   https://guix.gnu.org/en/blog/2021/gnu-guix-1.3.0-released
>

Cheers!  Thank you, to everyone who make it possible!

Re: Avoiding PYTHONPATH - latest?

[宋文武]

Hello!

Phil  writes:
>> I've been having an argument with myself over the last 4 days about if
>> Guix's use of PYTHONPATH is a necessary evil or avoidable on a foreign OS.
>>
>> I've found references to a similar discussion last year, and reference
>> to using a 'fake virtual environment' (does anyone have a reference to
>> the other thread referenced where fake venvs are demonstrated as not
>> working):

I remembered did that too, I think it works in most cases.


>>
>> https://lists.gnu.org/archive/html/guix-devel/2019-06/msg00204.html
>> https://lists.gnu.org/archive/html/guix-devel/2019-06/msg00221.html
>>
>> I had what I think is a similar idea over the weekend and tried it out.
>>
>> Reading the mechanics of venvs as per PEP:
>> https://www.python.org/dev/peps/pep-0405
>>
>> I decided to see if I added a fake pyvenv.cfg to the python.scm package
>> could allow for us to drop the use of PYTHONPATH completely.

I had came up with adding another environment variable
‘GUIX_PYTHON_X_Y_SITE_PACKAGES’:
  

But didn’t made it in, feel free take it if useful, thanks!

Re: Environment variables on GNOME on foreign distro (Debian)

[宋文武]

Jorge  writes:

> Hi.  On a previous thread [1] I asked about best practices for Guix
> environment variables on foreign distro.  My problem was not
> fundamentally solved, and it resurfaced again recently.

Hello, I'd say it's not a solved problem in general.

> In summary:
>
> I use Guix on an updated Debian buster (with 59 packages from
> buster-backports) to get some up-to-date packages atop Debian stable.
> Currently I set Guix environment variables on ~/.profile, but recently I
> started getting error messages from Evince (my Evince is from Debian's
> APT).  In fact, in /var/log/user.log I get:
>
> Feb 23 17:04:19 jorge--inspiron-5570 org.gnome.Evince.desktop[1788]: 
> /home/jorge/.guix-profile/lib/gio/modules/libdconfsettings.so: cannot open 
> shared object file: Permission denied
> Feb 23 17:04:19 jorge--inspiron-5570 org.gnome.Evince.desktop[1788]: Failed 
> to load module: /home/jorge/.guix-profile/lib/gio/modules/libdconfsettings.so
> Feb 23 17:04:19 jorge--inspiron-5570 evince[5252]: Using the 'memory' 
> GSettings backend.  Your settings will not be saved or shared with other 
> applications.
>
> The problem seems to be that Debian's Evince is seeing the
> `GIO_EXTRA_MODULES' environment variable from Guix.  In fact:
>
> $ sudo tr \0 \n < "/proc/$(pgrep evince)/environ" | grep GIO_EXTRA_MODULES
> GIO_EXTRA_MODULES=/home/jorge/.guix-profile/lib/gio/modules

Yes, the host evince (and any GNOME apps, in fact) is not compatible
with gio modules from guix, as they may have different glib versions and
ABI.

>
> Side note: `sudo` was unnecessary in that command.
>
> Ludovic Courtès on 12 Mar 2018 had suggested me to to source
>  ~/.guix-profile/etc/profile from ~/.bash_profile (or similar).
> However, that does not work, at least for me.  If I source Guix's
> etc/profile from ~/.bash_profile, the icons of my graphical Guix
> applications fail to appear in my GNOME docker, because XDG_DATA_DIRS
> was not set by Guix.
Well, ~/.profile and ~/.bash_profile should both works for bash, but
your login manager may not source ~/.bash_profile.

> In fact, if I open gnome-terminal and issue
> `env | grep -i guix`, I see no Guix environment variables.  In fact,
> that happens even inside Guix-installed emacs-next.

For ~/.profile (or ~/.bash_profile) to be loaded, gnome-terminal need to
launch a login shell (by default it's not).

>
> So what should I do?

I guess you can:
  figure out what environment variables from guix are needed, what are not.
(eg: filter out GIO_EXTRA_MOUDLES)
  figure out how the login manager got environment variables, and set them.
make sure XDG_DATA_DIRS from guix is added here, so that
applications can be found by the gnome launcher.
(eg: only from ~/.profile, or other?)
  for guix only environment variables, find a way to launch them.
(eg: when launch a guix gnome application that requires GIO_EXTRA_MODULES)

Hope it helps!

Re: Failed to autoload make-page-map in (charting)

[宋文武]

刘力铭  writes:

>> Note that mirror.hydra.gnu.org is defunct.  You should use
>> ci.guix.gnu.org, which is the only build farm that is provided by the
>> Guix project.
>
> Sadly, my network provider block ci.guix.gnu.org (I'm in mainland China)...
> Is there a mirror for ci.guix.gnu.org? I just found some mirrors for
> hydra.

It's not blocked for me (I have CMCC), and substitutes works, but the
url (a web ui for our CI) is 502 now.

>
>> This is unrelated to compiling code that can use Guile Charting at
>> runtime.  The warning is harmless.
>
> But this warning occurs as an error and block the installation process.

I think the real errors are at end of the message, could you check
again?

Re: Xfce4 localization in Guix System for only one exact user

[宋文武]

zna...@disroot.org writes:

> Yes, Julien, you are right!
> Changing config file to this:
>
> (operating-system ...
>   (locale "ru_RU.utf8"))
>
> and running as root:
>
> # guix pull
> # guix package -u
> # guix system reconfigure /etc/config.scm
> # guix package -u
>
> made all my applications work on Russian: Xfce, Gimp, LibreOffice, Audacity 
> and so on:
>
> http://0x0.st/zOsk.png

Yes, this changes the "LANG" system wide, in /etc/environment.

>
> But the next question is: how to make Russian environment only for defined 
> user, not for all users?
>
> You said, my DM may have it. I use Slim and did not found such
> configs. Also `guix search xfce` did not display something like
> xfce4-settings. And I did not found language settings in xfce settings
> manager.
>
> My question rests unanswered.

You can create a ‘.xsession’ file in that user’s home directory, with
something like:

#!/bin/sh
export LANG=ru_RU.utf8
exec "$@"

And make it executable by "chmod +x ~/.xsession".  Then when that user
login, the ‘.xsession’ will be executed.  Hope this helps!

Re: Installation of Guix hangs

[宋文武]

Tomasz Chrzczonowicz  writes:

> Hello everyone,
>
> I'm trying to install Guix 1.0.1.i686 on my old 1,6GHz, 1 GB RAM netbook.
>
> The installation hags during the "copying to /mnt" part.  Is there a
> way for me to troubleshoot the installation?

Hell, did you use the graphical installer?  Maybe you can switch the
console by "Ctrl-Alt-F2" etc?

>
> Is my hardware too old for the installation to work?

I think it's fine, but 1G RAM may not be enough, you can try setup a
swap first.

Re: A question about manifest for my profile

[宋文武]

hub.lomb...@free.fr writes:

> Hi Guix!
>
> I generated a 'manifest-to-manifest.scm' manifest of all the packages 
> installed in my user profile, using this Guile script (I can not find the 
> address anymore)
>
> 
> ;; Run with:
> ;; guile -s FILE ~/.guix-profile
>
> (use-modules (guix profiles)
>(ice-9 match)
>(ice-9 pretty-print))
>
> (define (guix-manifest where)
>   (sort (map (lambda (entry)
>(let ((out (manifest-entry-output entry)))
>  (if (string= out "out")
>  (manifest-entry-name entry)
>  (format #f "~a:~a"
>  (manifest-entry-name entry)
>  (manifest-entry-output entry)
>  (manifest-entries (profile-manifest where)))
>   string
> ;; Thanks to Ivan Vilata-i-Balaguer for this:
> (define (guix-commit)
>   (let ((guix-manifest (profile-manifest (string-append (getenv "HOME") 
> "/.config/guix/current"
> (match (assq 'source (manifest-entry-properties (car (manifest-entries 
> guix-manifest
>   (('source ('repository ('version 0) _ _
>('commit commit) _ ...))
>commit)
>   (_ #f
>
> (match (command-line)
>   ((_ where)
>(format #t ";; commit: ~a\n" (guix-commit))
>(pretty-print
> `(specifications->manifest
>   ',(guix-manifest where
>   (_ (error "Please provide the path to a Guix profile.")))
> 
>
> Then I called it:
>
> hubert@gnu ~$ guile -s manifest-to-manifest.scm ~/.guix-profile
> ;;; note: auto-compilation is enabled, set GUILE_AUTO_COMPILE=0
> ;;;   or pass the --no-auto-compile argument to disable.
> ;;; compiling /home/hubert/manifest-to-manifest.scm
> ;;; compiled 
> /home/hubert/.cache/guile/ccache/2.2-LE-8-3.A/home/hubert/manifest-to-manifest.scm.go
> ;; commit: 6c83c48391ebc2b08ca4704c437de4f1e59fca9a
> (specifications->manifest
>   '("brasero"
> "ffmpeg"
> "gimp"
> "gnumeric"
> "inkscape"
> "lynx"
> "markdown"
> "mpv"
> "obs"
> "qemu"
> "quassel"
> "racket"
> "sbcl-next"
> "ungoogled-chromium"
> "wget"
> "youtube-dl"
> "youtube-viewer"))
> hubert@gnu ~$
>
> Then I made the updates, reconfigured /etc/config.scm and restarted.
>
> I think the launch of the script was more or less equivalent to the following 
> command:
>
> $ guix package --manifest=manifest-to-manifest.scm

This '--manifest' option accept a file, which should return a manifest
object, and the script you run generate it from your profile.  So you
can write the script output "(specifications->manifest .)" into a
file, eg: "~/.config/guix/manifest.scm", then use "guix package -m
~/.config/guix-manifest.scm" to apply it.


>
> In which case, I will not need to run this command?
>
> And finally, are there any changes to make in '/etc/config.scm'? Maybe add to 
> it:
>
> ---
> (specifications->manifest
>   '("brasero"
> "ffmpeg"
> "gimp"
> "gnumeric"
> "inkscape"
> "lynx"
> "markdown"
> "mpv"
> "obs"
> "qemu"
> "quassel"
> "racket"
> "ungoogled-chromium"
> "wget"
> "youtube-dl"
> "youtube-viewer"))
> ---


No, the "packages" field of "operation-system" accept a list of package
objects, it can't deal with "manifest".  In fact, "manifest" is used to
manage user's own profile, in a declarative way, just like "packages".

>
> Thank you in advance for your possible clarification :-)

Yeah, hope it helps!

Re: Writing and Bulding emacs-weechat package: In procedure url-fetch, Invalid keyword

[宋文武]

b0f0  writes:

> Hello Guix !
> (This is my first mail to a newsgroup. )

Welcome!
(This is a mailing list, a newsgroup should be accessed via NNTP instead.)

> I would like to package the emacs-weechat package.
> I run this command in my terminal:
>
> $guix import elpa -a melpa emacs-weechat
>
> I put the output in to a emacs-weechat.scm file
> and then I modified the output like this (everything below this line
> is a emacs-weechat.scm):
>
> ;;(define-module (gnu packages emacs-weechat)
> ;;   #:use-module (guix packages)
> ;;   #:use-module (guix build download)
> ;;   #:use-module (guix build-system emacs)
> ;;   #:use-module (gnu packages emacs-xyz))
> (use-modules (guix packages)
>(guix download)
>(guix git-download)
>(guix utils)
>  (guix build download)

You don't need ‘(guix build download)’, and it exports another
‘url-fetch’ proceduce which cause errors here…

>(guix build-system emacs)
>(gnu packages emacs-xyz))
>
> (package
>   (name "emacs-weechat")
>   (version "20190520.1551")
>   (source
> (origin
>   (method url-fetch)
>   (uri (string-append
>"http://melpa.org/packages/weechat-";
>version
>".tar"))
>   (sha256
>   (base32
> "143bjyrfz37ihmpkrbr668pn0z1017pjzb1d7hmqrr518slg7a"
>   (build-system emacs-build-system)
>   (propagated-inputs
> `(("emacs-s" ,emacs-s)
>   ("emacs-tracking" ,emacs-tracking)))
>   (home-page "https://github.com/the-kenny/weechat.el";)
>   (synopsis "Chat via WeeChat's relay protocol in Emacs")
>   (description "This package provides a way to chat via WeeChat's
> relay protocol in Emacs. Please see README.org on how to use it.")
>   (license #f))
>
>
> I read the manual but I am stuck, in terminal I run this comand:
> $guix build --file=emacs-weechat.scm
>
> Backtrace:
> In guix/store.scm:
>623:10 19 (call-with-store _)
> In guix/scripts/build.scm:
>911:26 18 (_ #)
> In ice-9/boot-9.scm:
> 829:9 17 (catch _ _ # …)
> In guix/ui.scm:
> 495:6 16 (_)
> In guix/scripts/build.scm:
> 876:5 15 (_)
> In srfi/srfi-1.scm:
>679:15 14 (append-map _ _ . _)
>592:17 13 (map1 ("x86_64-linux"))
>679:15 12 (append-map _ _ . _)
>592:17 11 (map1 (#))
> In guix/scripts/build.scm:
>835:18 10 (_ _)
> In guix/packages.scm:
>936:16  9 (cache! # # …)
>   1255:22  8 (thunk)
>   1188:25  7 (bag->derivation # #<…> …)
> In srfi/srfi-1.scm:
>592:29  6 (map1 (("emacs" #) …))
>592:17  5 (map1 (("source" #https://melpa.org/packa…>) …))
> In ice-9/boot-9.scm:
> 829:9  4 (catch srfi-34 # …)
> In guix/packages.scm:
>   1003:18  3 (_)
> In guix/store.scm:
>   1794:24  2 (run-with-store _ _ #:guile-for-build _ #:system _ # _)
>   1667:13  1 (_ _)
> In guix/build/download.scm:
> 741:0  0 (url-fetch _ _ #:timeout _ #:verify-certificate? _ # _ # …)
>
> guix/build/download.scm:741:0: In procedure url-fetch:
> Invalid keyword: #vu8(234 60 170 81 40 57 227 10 79 11 235 203 123 2 8
> 31 216 139 140 201 43 207 91 97 60 227 187 236 229 26 36)
>
> Isn't url-fetch provided with 'guix download' module ?

Yes, and remove ‘(guix build download)’ does the right thing for me.

Have a good day :-)

Re: GNU Guix 1.0.0 released

[宋文武]

Ludovic Courtès  writes:

> We are thrilled to announce the release of GNU Guix 1.0.0!
>
> This 1.0 release is a major milestone for Guix.  It represents 7 years
> of hard work with more than 40,000 commits by 260 people, 19 releases,
> and an equally amazing amount of work on documentation, translation,
> artwork, web design, mentoring, outreach, and many other activities that
> together have made it a thriving project.
>
> Read more about today’s announcement at:
>
>   https://gnu.org/software/guix/blog/2019/gnu-guix-1.0.0-released
>
> Whether you’re a software developer, a user, or a free software
> enthusiast, we hope GNU Guix will provide you with the tools to deploy
> and manage software with confidence and ease, qualities that are not
> usually associated with software deployment.  We’d love to hear from you!
>

Congratulations for the great 1.0.0 release, and thank you very much!!

And for all ones that sending/reviewing/applying patches, packaging
softwares, making system services, writing and translating
documentations, helping each other, hacking for the good, so anyone
involved, thank you!

GNU Guix is such an excellent project and community, I learned a lot and
gain much satisfactions with it, and will continue to.

See you here in the next 7 years!

Re: xorg-configuration questions

[宋文武]

John Soo  writes:

> Hey all,
>
> I love the idea of having one record for keyboard layout and I just 
> reconfigured my system with the new configuration but my settings won't take. 
> I've had issues with the old
> `extra-config` options in the past (nothing worked then including the 
> cst-trackball from below). Admittedly, I am quite unfamiliar with xorg 
> configuration of input devices so I can
> imagine I may be doing something wrong. I've listed what I have so far below. 
> Note that the console keymap works properly. Also the auto-login does not 
> seem to work either.
> What am I doing wrong? 
>
> (define ctrl-nocaps (keyboard-layout "us" #:options '("ctrl:nocaps")) 
> ...
> (define cst-trackball
>  "Section \"InputClass\"
> Identifier \"CST Trackball\"
> Driver \"libinput\"
> MatchVendor \"CST\"
> MatchProduct \"CST USB UNITRAC\"
> MatchIsPointer \"on\"
> Option \"AccelerationNumerator\" \"2.0\" EndSection")
> ...

Hello, I think there is no "AccelerationNumerator" option for libinput, it
has "AccelSpeed" instead:

https://wiki.archlinux.org/index.php/Mouse_acceleration#Mouse_acceleration_with_libinput

Hope it helps!

Re: Does not halt after Menu - Logout - Shut down and closing notebook hood

[宋文武]

 writes:

> When I return to notebook (even 8-10 hours later) and open it I see it is 
> working and lines in a black screen go there. Something like 'stoping 
> service.. will now halt'. And it turning off.
>
> Please, how to configure it not to sleep when I close hood but to halt when I 
> want it this way?
>
> Hello, Guix Help! I want to make an effort to solve a problem I've noticed.
> I am using GuixSD, Xfce4. I think I have no screensaver and Screen lock.
> But inconvenient thing rests there.
> Every time I want to turn of notebook and close it quickly it does not switch 
> off at all.
> 1. I click Xfce4 Menu -> Log out -> Shut down
> 2. I close notebook hood quickly
> 3. I go away
> 4. Notebook keep on working
>
> When I return to notebook (even 8-10 hours later) and open it I see it is 
> working and lines in a black screen go there. Something like 'stoping 
> service.. will now halt'. And it turning
> off.
>
> Please, how to configure it not to sleep when I close hood but to halt when I 
> want it this way?

Hello, I think this can be done by set 'handle-lid-switch' to 'ignore'
(it's 'suspend' by default) for the 'elogind-service-type', for example:

(operating-system
  ...
  (services
(cons* (xfce-desktop-service)
   (modify-services %desktop-services
 (elogind-service-type
   c => (elogind-configuration
  (handle-lid-switch 'ignore)))

Hope it helps!

Re: racket

[宋文武]

Benjamin Slade  writes:

> When run Racket, in the REPL, the first command I execute runs fine. But
> the second (no matter what it is, even if it's just the first again)
> always produces:
>
> ; ptr-ref: contract violation
> ;   expected: (and/c cpointer? (not/c (lambda (p) (pointer-equal? p #f
> ;   given: #f
> ;   argument position: 1st
> ; [,bt for context]
> ; ptr-ref: contract violation
> ;   expected: (and/c cpointer? (not/c (lambda (p) (pointer-equal? p #f
> ;   given: #f
> ;   argument position: 1st
> ; [,bt for context]
>
> And it doesn't accept any more input after that point (including ",bt").

Hello, I fixed it with commit 26700cae.  A 'guix pull' and then install
racket again should work.

Thank you for the report!

Re: Something broke

[宋文武]

Jone  writes:

Hello!

> Hi, people! There are also prepared packages - why not a give to the
> world this? As third-patry, no warranty.
> Or is it still an experimental distro for hackers? I have prepared 3
> packages, they are not in the official delivery, but where can I send
> them? 

If you have made new guix packages, you can send patches to
guix-patc...@gnu.org, See the "Submitting Patches" section in:

https://www.gnu.org/software/guix/manual/html_node/Contributing.html

Does this help?  Thank you!

Re: Multilib support: running 32-bit binaries on 64-bit systems

[宋文武]

Pierre Neidhardt  writes:

> Guix already supports cross-compilation out of the box of 32-bit
> packages on any architecture:
>
>   guix build --system=i686-linux PACKAGES...
>
> Now the question is: is it possible to run the produced binaries
> directly (no VM)?

Yes, you can run i686 binaries directly on a x86_64 system with the default 
libre
linux kernel (which have i686 support enabled), one example is the
‘wine’ package.

>
> Many other distributions have support for the so-called "multilibs",
> that is, a separate 32-bit library tree with the 32-bit linker
> configured to load those instead of their 64-bit counterparts.
>
> Does Guix have any provision for multilibs?

Guix install every package into its own directory which already cover
the multilibs cases.

Re: Error system reconfigure with guix-latest

[宋文武]

Jone  writes:

> Hello!
> I'm newbie. After update system (guix pull) I see this error:
>
> root@guix ~# guix system -n reconfigure /etc/CURRENT.scm
> /etc/CURRENT.scm:37:24: error: you need these modules in the initrd for
> /dev/sdb2: sata_nv pata_acpi

Hello, this is due to recently added functinon to detect missing modules
in the initrd: .

To fix it, you can use 'initrd-modules' instead of the 'initrd' field in
the 'operating-system' record.

>
> BUT:
> 1. Prior to this (before upgrade) everything was OK
> 2. sata_nv and pata_acpi already selected in CURRENT.scm:
>
> (operating-system
> (host-name "guix")
> (timezone "Europe/Moscow")
> (locale "en_US.utf8")
> (bootloader (bootloader-configuration
> (bootloader grub-bootloader)
> (target "/dev/sdd")))
> (initrd (lambda (file-systems . rest)
> (apply base-initrd file-systems
> #:extra-modules '("sata_nv" "pata_acpi")
> rest)))

This '#:extra-modules' usage is deprecated by the newly added
'initrd-modules' field in 'operating-system'.  For example:


> (file-systems (cons*
> (file-system
> (device "two")
> (mount-point "/")
> (type "ext4")
> (title 'label))
> (file-system
> (device "one")
> (mount-point "/home")
> (type "ext4")
> (title 'label))
> %base-file-systems))
>
> I didn't find any answers in GuixSD documentation..

Well, the guix info manual did got updated, but currently 'guix pull'
won't bring it, for reference: .

Re: Performance issues with /gnu/store in virtual machines

[宋文武]

l...@gnu.org (Ludovic Courtès) writes:

> Hi Chris,
>
> Christopher Baines  skribis:
>
>>   1.130.19   0  3853   openat
>
> [...]
>
>>  13.560.193014  50  3853   openat
>
> So openat(2) is 4 orders of magnitude slower in the VM, which is a
> problem.
>
> I guess a large part of that is due to the fact that VM created with
> ‘guix system vm’ use unionfs as their root file system, which is a FUSE
> thing (see the “volatile” options in (gnu system vm), (gnu system
> linux-initrd), and (gnu build linux-boot).)  I’m not sure how we could
> avoid it.

We might switch to the linux kernel's bultin overlayfs, which should be
faster.

https://wiki.archlinux.org/index.php/Overlay_filesystem

Re: retroarch on Trisquel

[宋文武]

Caleb Herbert  writes:

> Why does retroarch have black rectangles where icons should be, and
> missing paths to fonts?
>
> Why does the file browser not detect z64 files?

I think these are due to missing assets and core info files, which can
be downloaded via the 'Online Updater'.  Hope it helps!

Re: StumpWM package doesn't work

[宋文武]

Toni Reina  writes:

> Hello,
>
> I'm trying to install `sbcl-stumpwm` package and looks like it doesn't
> work correctly. It's installed with no errors, but it doesn't generate
> the stumpwm binary file.
>
> The following package will be installed:
>sbcl-stumpwm 0.9.9   
> /gnu/store/z92ri0kgjdavkp7llav1db0dia44sbid-sbcl-stumpwm-0.9.9
>
> ls /gnu/store/z92ri0kgjdavkp7llav1db0dia44sbid-sbcl-stumpwm-0.9.9
> -> lib  share
>

It's in the "bin" output of sbcl-stumpwm package, you can get it with:

guix package -i sbcl-stumpwm:bin

Re: Some questions regarding GuixSD

[宋文武]

dian_ce...@zoho.com writes:

> On Sat, Oct 01, 2016 at 09:50:11AM +0800, 宋文武 wrote:
>> dian_ce...@zoho.com writes:
>> 
>> On GuixSD, you can rollback the whole system safely and easily (by
>> choosing a grub menu item at boot).
>
> I don't guess you happen to be able to supply me with the options passed to do
> this, would you?

Ah, that's managed by guix on GuixSD, which produces a grub.cfg with menu
containing all systems (current and previous if not deleted explicitly)
it ever built.  No option needed.

> My systems aren't exactly standard installs, and I don't use
> GRUB for anything at the moment (this may or may not change depending on 
> where I
> feel Guix fits into my systems, and whether I end up migrating everything 
> over,
> or if it's just one computer that has a GuixSD install on it).

OK, muli-boot with GuixSD can be done with chainloader, etc. but it's a
bit tricky.  You probably want to try it in a VM (eg: guix system vm)
first :-)

Re: Some questions regarding GuixSD

[宋文武]

dian_ce...@zoho.com writes:

> On Fri, Sep 30, 2016 at 10:18:40AM +0200, Ludovic Courtès wrote:
>> > 2) Are multiple versions of a package in the tree at any given time?
>> 
>> You can have multiple profiles, each containing a different version of a
>> given package, for example.
>> 
>> The source tree of Guix itself usually only contains the latest version
>> of each package.  There are exceptions for packages where it makes sense
>> to keep several stable series in parallel: GCC, Python, Guile, etc.
>
> I don't know where Guix/GuixSD is in regards to the plans for it, and I havn't
> had a chance to glance at the ROADMAP you mentioned later in this email, but I
> can say from personal experience that maintaining one and only one version of 
> a
> package in a tree is a Really Bad Idea. When I used Arch they had (and may 
> still
> have, I have no clue what has happened with that distro since) a policy to 
> only
> maintain one version of any package at any given time in the distro. This is a
> huge PITA and sometimes can make a system unusable if a package breaks. I
> distinctly recall having one package I relied on breaking when I used Arch and
> having to wait a few days for them to release a new version.

On GuixSD, you can rollback the whole system safely and easily (by
choosing a grub menu item at boot).  And the CI service (hydra) is
building (each package if any of its inputs changed) and testing (system
with services in qemu) the whole distribution.

While in Guix it's easier than Debian or ArchLinux to have multiple
versions of the same packages, it's reasonble only do it when needed :-)

>
>> 
>> > 3) Is there any way to install GuixSD from something other than the LiveUSB
>> > image on the site?
>> 
>> You can install Guix on top of your GNU/Linux system and run:
>> 
>>   guix system init config.scm /
>> 
>> Be careful: this is a one-way change!
>
> So as long as Guix is installed correctly in the host system, the 
> aforementioned
> command is effectivly the install phase, while the rest is simply setting up 
> the
> enviroment for the new system? I can certain get behind that.

Yes, the install phase mainly contains:

- build the system from config, only write to /gnu/store. (safe)
- deploy the system to target, overwrite /etc, /var/, etc.
  and install the grub bootloader. (dangerous)

You can install it into a seperate partion (same as in the Manual),
it's more safe IMO.

Re: .xinitrc ignored

[宋文武]

gno  writes:

> Hello guys !
>
> First off, guix is great, even if it takes quite some time to get used
> to :)
Yeah, welcome!

>
> I made it work in no time and am now running guixsd (lightweight
> desktop setup with awesome). I'd like to keep at it but I do need to
> solve 2 problems first. I was asking around in #guix and was referred
> to this email add.
>
> Is there any way to make it respect .xinitrc, preferrably when slim
> starts ? I use a diy keyboard with a custom xmodmap that needs to be
> loaded. I tried all kinds of things with ~ and
> root; .xinitrc, .xserverrc, .xsession (crashed w/o logs).
Our slim only supports `.xsession', which can be a shell script with
executable permission.  It accepts the session command selected by
the user, and should exec it (or something).  My is look like:

--8<---cut here---start->8---
#!/bin/sh

setxkbmap dvorak
$HOME/.fehbg
xrdb ~/.Xresources

# exec "$1" # uncomment this line to start the selected one 
exec dbus-launch --exit-with-session i3
--8<---cut here---end--->8---

>
> Also, unfortunately awesome-wm is quite dated. I'd like to update it so
> that it is compatible with my config. There have been quite a few API
> changes. Is it feasible to try and package a newer version myself or
> did you not do it already because of missing dependencies or other
> complications ?

Yes, the most important missing one is lua-lgi.  And we don't have
other lua packages too, so I think it's most about figure out
how lua packages should works (finding them during applications
build and running) in guix.

Re: How to install an old package version?

[宋文武]

Hartmut Goebel  writes:

> Hi,
>
> I'm curious about how to install an old package version with guix.
>
> Example for what I mean:
>
> In e.g Debian, the list of available packages is separate from apt-get
> (et al.). So I can query all available versions of a package and
> install the version I need:
>
> $ apt-cache madison nginx
> nginx | 1.9.10-1~bpo8+3 | http://debian.mirror.lrz.de/debian/
> jessie-backports/ma
> nginx | 1.6.2-5+deb8u2 | http://security.debian.org/
> jessie/updates/main amd64 Pa
>
> Now in guix, the list of available packages is build into guix, there
> is no external cache. So how can I e.g. install python-2.7.10 after I
> installed guix 0.11.0, which only defines python-2.7.11?

All definitions (like debian's dsc) of packages are in the guix
git reposiroy.  You can clone it, checkout to any commit, build it,
and then run `pre-inst-env guix package -i ...' there to install
packages at that commit.  Note that guix hydra only has binary caches
for latest builds, so the old packages will be built from sources.

You can also 'guix pull --url=...' to update (or downgrade) the
package definitions from a tarball, which can be generated from the
git repository by `git archive'.

Re: uwsgi and Pyramid

[宋文武]

Myles English  writes:

> Hello,
Welcome!

>
> I have defined some packages to install Pyramid (a Python web
> framework) here:
> https://gist.github.com/mylese/861b4765e9f533bafa2cb2b305974caa
>
> Now I am trying to get my python app (my_app), served by uwsgi
> but get the infamous "no app loaded" error from uwsgi:
>
>
> $ /gnu/store/y83n9b6zl76pq1v7qb0npl7p7gd0hgd7-profile/bin/uwsgi --need-app 
> --ini production.ini

It looks to me that this production.ini file is for paste.deploy, so you
need add 'python-pastedeploy' to the environment and pass '--ini-paste'
instead of '--ini' to uwsgi.

Re: inputs vs. native-inputs vs. propagated-inputs

[宋文武]

Hartmut Goebel  writes:

> Hi,
>
> I'm a bit confused about the difference between inputs, native-inputs
> and propagated-inputs.
>
> The manual states:
>
>   The distinction between ‘native-inputs’ and ‘inputs’ is
>   necessary when considering cross-compilation.  When
>   cross-compiling, dependencies listed in ‘inputs’ are built for
>   the _target_ architecture; conversely, dependencies listed in
>   ‘native-inputs’ are built for the architecture of the _build_
>   machine.
Yes, _native_ means the same architecture as the build machine.

>
> For for I understand. But then the manual says:
>
>   ‘native-inputs’ is typically used to list tools needed at
>   build time, but not at run time, such as Autoconf, Automake,
>   pkg-config, Gettext, or Bison.
>
> The first sentence implies that "inputs" are treated as needed at run
> time.
No, as _native_ inputs usually are tools for building (and testing),
most time they’re not needed at run time.

The dependencies of a store item (an output of the package) is
computed by scan all its files for store paths (as reported
by ‘guix gc –-references …’).  When you download a store item by
substitutes, its dependencise will be downloaded too.  When you
install a store item by ‘guix package -i /gnu/store/…’, only one
path will be added to the profile.


The runtime dependencies are only for _package_ objects.
When you add a _package_ object into profile (or inputs for
building other package), its ‘propagated-inputs’ will be added too.


>
> Now consider libAAA (supporting the famous Amiga AAA chipset ;-), which
> requires the headers of libBBB to compile (but only to fetch some
> constant definitions) and libCCC at run-time. Without libCCC, libAAA
> could not work. And libAAA uses pkg-config to find the header files.
>
> So for me this would be:
>   libBBB: inputs
>   libCCC: propagates inputs
>   pkg-confg: native inputs
>
>
> Is this correct?
No, libCCC not necessary be propagated, it’s a dependency of libAAA,
but this dependency maybe transparent if the use of libAAA doesn’t
requires libAAA.

- To build libAAA natively, be inputs is enougth.
- To build libAAA for other architecture, pkg-config needed to
  be ‘native-inputs’, due to it’s an ELF executable to be invoked by the
  build system.
- To make the users of libAAA happy, libBBB or libCCC should be propagated
  if the use of libAAA require them (eg: its headers include libBBB’s
  headers or its .pc file list libCCC in the ‘Requires’ field.).


>
> If so, how can I as a packager find out if eg. libBBB is only used at
> build time and libCCC need to be a propagated input?
By looking the output of ‘guix gc –-references …’, the build time ones
are ones not there.  If you think something shouldn’t be there (eg:
build a library, but gcc end up within its references), then it’s time
to patch it to reduce closure size.

For propagated inputs, it’s for the users.  So if (for programs) we run
it fine or (for libraries) include its headers and link it fine, nothing
need to be propagated.

>
> Same for pkg-config: How to determine if this is only needed ar build
> time (as I would always expect)? The manual says:
>
>   … propagated-inputs …
>   For example this is necessary when a C/C++ library needs
>   headers of another library to compile, or when a pkg-config
>   file refers to another one via its ‘Requires’ field.
>
>   For me this is confusing.
Many build systems use ‘pkg-config’ to check for libraries and get
flags, a pc file usually list some other packages in its ‘Requires’
field, if one of these packages is missing (doesn’t have a .pc
file in PKG_CONFIG_PATH), this pc file will be treated as broken, and
the package will be reported as ‘not found’.  So propageted these
packages make ‘pkg-config’ works, reduce the work for packaging its
users (otherwise, those packages need to added as inputs even they’re
not used directly).

Re: icedtea is installed, but no javac available.

[宋文武]

Dmitry Nikolaev  writes:

> Hi. I have icedtead both in my /etc/config.scm and in my user profile:
>
> $ guix package --list-installed | grep icedtea
> icedtea 3.0.1 out
^ This means the "out" output.
> /gnu/store/krksxq6wzxi4bj9vc7g02axgjb7nrmgi-icedtea-3.0.1
>
> $ ls /gnu/store/ | grep -e '.*icedtea.*-jdk'
> 1rxafkp02y1775viisji16nfam7syqq3-icedtea-2.6.6-jdk/
> 90aj0p2yfpcizf3dl0bds89i9ggx9pbq-icedtea-3.0.1-jdk/
> 9ls43rishmxbc9bwyl6v55a3rpsjdrqz-icedtea-3.0.1-jdk/
> ivlpx5z72r2b2gzbx2fzvk43f3q1hq70-icedtea-2.6.5-jdk/
> k9b628hrxrxyyn4mn6wsvpy530w9mrv2-icedtea-2.6.6-jdk/
> qcgmxr29w967y8xzavls6zs1bcw6d8md-icedtea-2.6.5-jdk/
 ^ This is the "jdk" output.
>
> $ ls -l
> /gnu/store/ivlpx5z72r2b2gzbx2fzvk43f3q1hq70-icedtea-2.6.5-jdk/bin/javac
> -r-xr-xr-x 2 root guixbuild 7816 Jan 1 1970
> /gnu/store/ivlpx5z72r2b2gzbx2fzvk43f3q1hq70-icedtea-2.6.5-jdk/bin/javac
>
> $
> /gnu/store/ivlpx5z72r2b2gzbx2fzvk43f3q1hq70-icedtea-2.6.5-jdk/bin/javac
> -version
> javac 1.7.0_99
>
> But javac is not in my $PATH, neither in root's $PATH:

You installed the "out" output of icedtea, which is only the JRE part.
To install the "jdk" output, use: `guix package -i icedtea:jdk'.

Re: How to remove old (test) builds from store?

[宋文武]

Hartmut Goebel  writes:

> Hi,
>
> when creating packages and refining the build step by step, I end up
> with many entries in the store I do not need. When tying to delete them,
> only a few are removed and many are kept. E.g:
>
>
> # guix gc -d /gnu/store/*teensy*
> finding garbage collector roots...
> deleting `/gnu/store/…-teensy-loader-cli-2.1-1.f289b7a.drv'
> deleting `/gnu/store/…-teensy-loader-cli-2.1-1.f289b7a.tar.xz.drv'
> guix gc: error: build failed: cannot delete path
> `/gnu/store/…-teensy-loader-cli-2.1-1.f289b7a.tar.gz.drv' since it is
> still alive
> $ ls -d /gnu/store/*teensy* | wc -l
> 35
>
> The one still alive if okay, since I have an older version of this
> package installed in my current environment.
>
>
> How do I get rid of these ca. 30 outdated store items?
I think call gc for each one will work, eg:

for i in /gnu/store/*teensy*; do guix gc -d $i; done

Re: configuring font rendering in GuixSD

[宋文武]

Stu Kraji  writes:

> hello. 
>
> Usually one can configure font rendering with symlinks in some
> directory like /etc/fontconfig/conf.d or using a configuring file. 
>
> I tried the directory ~/.config/fontconfig but it does not work.
>
Hi!

While the system-side config is read-only in /gnu/store,
the user-side ~/.config/fontconfig does work for me.

According to ‘50-user.conf’, it can be:
  ~/.config/fontconfig/fonts.conf (file)
or  
  ~/.config/fontconfig/conf.d (directory)

I have ‘~/.config/fontconfig/conf.d/00-all.conf’,
maybe you have ‘~/.config/fontconfig/zz-xxx.conf’?

Re: Typing on GuixSD in languages other than English

[宋文武]

iyzs...@member.fsf.org (宋文武) writes:

> Chris Marusich  writes:
>
>> Awesome!  Where you able to get this working?  I've tried installing it
>> From the version you pushed, and I still can't get it to work.  Even
>> when I set all the environment variables mentioned, and I install all
>> the packages (ibus, ibus-anthy, and anthy) to my profile.  If you got it
>> working, I'd like to know how so I can set it up on my end, too.

Yes, I only tested in xterm.
set the environemnt variables, run ‘ibus-daemon –-xim -v’.
then use ‘ibus-setup’ to add ‘Japanese - Anthy’.
in xterm, ‘Super+Space’ work fine.


For GTK+ applications, I have to generate the ‘immodules.cache’ file:
--8<---cut here---start->8---
GTK_IM_MODULE_FILE=/tmp/immodules.cache gtk-query-immodules-3.0 -–update-cache

GTK_IM_MODULE_FILE=/tmp/immodules.cache gtk3-demo –-run search_entry
--8<---cut here---end--->8---

With the cache (it contains ibus), it will work.

Re: Typing on GuixSD in languages other than English

[宋文武]

Chris Marusich  writes:

> Luis Felipe López Acevedo  writes:
>
>>> Chris Marusich  writes:
>>>
>>> The attached patches add anthy and ibus-anthy.  They build OK.
>>> However,
>>> Japanese input is not yet working for me.  In fact, ibus itself does
>>> not
>>> seem to be working at all.
>>>
>>> How can I verify that ibus is working on its own?  I've tried
>>> installing
>>> ibus by itself into my profile.  When I do this, I can run ibus-setup,
>>> and I see a "Japanese" language input option.  This is true even when
>>> anthy and ibus-anthy are not installed, which surprised me. I had
>>> expected to find no Japanese language input options at all in
>>> ibus-setup
>>> when anthy and ibus-anthy are not installed.
>>>
>>> In GNOME, there is no ibus menu visible anywhere on my screen, so it is
>>> unclear whether ibus is actually working.  When I open Gedit and press
>>> +Space, my input method does not seem to change.  According to
>>> ibus-setup, this hotkey should change my input method.  However, it
>>> seems like ibus is just not working at all.
>>
>> FWIW, this is what I do on Debian 8 (with GNOME SHELL):
>>
>> - Install ibus, ibus-anthy (or ibus-mozc).
>> - Restart the system.
>> - Go to *System settings → Region & Language*.
>> - Add *Japanese (Anthy)* in Input method.
>
> "Japanese (Anthy)" does not show up as an input method.  There are a
> variety of entries (which are present regardless of whether ibus, anthy,
> or ibus-anthy are installed), and none of them mention "Anthy".
>
The exec path in the anthy.xml is wrong, also we need wrap them with
python-pygobject.  I fixed that and pushed.

Thanks!

Re: Starting ssh service

[宋文武]

Dick Middleton  writes:

> Hello,
Hi!
>
>   Just started playing with Guix and am trying to get ssh to start as a 
> service.
>
> My system is configured using the standard desktop example config.scm given
> with the installation.
>
> I'm trying to add the relevant bits from the simple example in reference guide
>   "7.2.1 Using the Configuration System".
>
> i.e I've added:
>
> (use-service-modules networking ssh)
> ...
> (services (cons* (lsh-service) %base-services))
The desktop example use '%desktop-services' (in gnu/services/desktop.scm).
It contain services like dbus, slim, and wicd, etc.

wicd-service is provide the 'networking'.
>
> guix reconfigure complains with this error: ssh-daemon requires 'networking'
> which is undefined.
>
> Do I have to install something (why is it not already installed?) or am I
> doing something wrong?
You can use '%desktop-services' (or add 'wicd-service' or others.
see the 7.2.7.2 Networking Services section of the manual).

Hope it helps!

Re: Typing on GuixSD in languages other than English

[宋文武]

Chris Marusich  writes:

> Hi,
>
> I'm using GuixSD v0.10.0 with GNOME.  I'd like to be able to type in
> Japanese.  I can already display Japanese text without issue because
> I've installed fonts that contain Japanese characters; however, I cannot
> figure out how to enable Japanese input.
>
> In the past, I've done this for other GNU/Linux distributions.  I recall
> that the process was a little complicated.  Has anyone set up Japanese
> language (or any other non-English) input for GuixSD?  I'm going to look
> into this regardless, but if someone has already jumped through the
> hoops, I'd love to hear about how you did it.
There're fcitx and ibus as input method frameworks, I think we just need
to package anthy, ibus-anthy and fcitx-anthy.

Re: How to solve /etc/login.defs

[宋文武]

Jean Louis  writes:

> Hello,
>
> I am using guix on Debian. The default path for guix is:
> $HOME/.guix-profile/bin and sbin
>
> And I am using login manager lightdm, so the default PATH cannot be
> read. I guess, that /etc/login.defs have to be changed for each user to
> include $PATH correctly.
>
> Otherwise, how is system supposed to know that default $PATH is
> ~/.guix-profile/bin? among others?
In GuixSD, this is done by `/etc/profile' which is read by sh-compatible
login shells:
--8<---cut here---start->8---
if [ -f "$HOME/.guix-profile/etc/profile" ]
then
  # Load the user profile's settings.
  GUIX_PROFILE="$HOME/.guix-profile" \
  . "$HOME/.guix-profile/etc/profile"
else
  # At least define this one so that basic things just work
  # when the user installs their first package.
  export PATH="$HOME/.guix-profile/bin:$PATH"
fi
--8<---cut here---end--->8---

Re: guix package -i emacs, failed

[宋文武]

Jean Louis  writes:

> Hello,
>
> Yesterday I was trying to compile emacs from guix: guix package -i
> emacs, and it did not work, it said: gdk-pixbuf could not be done for
> 3600 seconds, something like that.
I think the issue is gdk-pixbuf's stress tests try to use a lot of
memory and sometimes can run too long without any output.  There is
an 'max-silent-time' option in guix can be used to set the 3600s limit,
but it may be not what we need here ;-)
>
> However, when emacs is compiled directly, it works without problems.
>
> Maybe gdk-pixbuf should not be dependency for emacs. It is basic to have
> emacs first.
gdk-pixbuf is used by gtk+, and the default emacs is compiled with gtk+
GUI.  There is 'emacs-no-x' with ncurses UI which don't need it.

Re: Error while logging in

[宋文武]

Pietro Ferrera  writes:

> done that; three times, actually. ):
Do you use ~/.xsession file?  It need to be an executable.

Re: (dynamic-link "libm") doesn't work on guile

[宋文武]

Alex Vong  writes:

> Hi people,
>
>
> I got the following error message when trying to dlopen a shared library
> in ~/guix-profile/bin/guile installed by guix:
>
> scheme@(guile-user)> (dynamic-link "libm")
> ERROR: In procedure dynamic-link:
> ERROR: In procedure dynamic-link: file: "libm", message: "file not found"
>
> Entering a new prompt.  Type `,bt' for a backtrace or `,q' to continue.
>
>
> I think it is a guix-specific problem since /usr/bin/guile installed by
> debian works fine:
>
> scheme@(guile-user)> (dynamic-link "libm")
> $1 = #
The `dynamic-link` will search and dlopen the shared library.  The guile
from guix doesn’t search common places like ‘/usr/lib’, so you need to
set the environment variable `LD_LIBRARY_PATH` or `LTDL_LIBRARY_PATH`
explicitly to the directory contains `libm.so`.

by the way, since guile is linked with libm.so, we can just use
`(dynamic-link)` here.

well, I also find `(dynamic-link "libm-2.22")` work too, but it doesn’t
seem to be what we want to use.

Re: Gnome-shell metapackage

[宋文武]

"d.4.n.1"  writes:

> I try again and work it! gnome-shell on guixsd! I'll help to add more
> packages to it. thanks ludo, David and all guix devs.
That's great to know!
>
> Sent from my Mi phone
> On l...@gnu.org, Jan 7, 2016 11:15 AM wrote:
>
> Hello! 
> 
> d.4@riseup.net skribis: 
> 
> > do you think about create a metapackage to gnome-shell? 
> 
> I think 宋文武 is kinda looking into it (or should consider
> looking into 
> it). :-)
Yeah, comparing with the 'gnome-core' meta package from debian,
useful packages are still missing, or maybe we should just add what
we have now and complete it gradually?
> 
> > How do I change slim/wicd to gdm/networkmanager? 
> 
> There’s no GuixSD service for GDM yet, but that’s another thing we 
> should look into it soon.
GDM can't start the X server (it hardcode /usr/bin/X, etc.) now, my plan
is to add an 'xorg-service' for '/run/setuid-programs/X' and then patch
GDM to use that instead.
> 
> There is ‘network-manager-service’ in (gnu services networking).
> It can 
> be used as shown here: 
> 
> https://lists.gnu.org/archive/html/guix-devel/2015-11/msg00717.html
>
> 
> It didn’t work for me though, so your feedback would be
> appreciated! 
> 
> 宋文武, I noticed that ‘network-manager-service’ is undocumented
> in the 
> manual; could you add it under “Networking”?
OK, will do it.

Re: Configuration System's Handling of Missing Packages

[宋文武]

Keith Osterheld  writes:

> Hi,
>
> How does the configuration system handle references to packages that
> aren't installed? For example, if I add (use-package-modules xfce) to my 
> config file
> but I don't have xfce installed, will guix automatically install it, or
> will it throw errors?
“(use-package-modules xfce)” is same as “(use-modules gnu packages xfce)”,
it only make the ‘xfce’ module avaliable (like ‘import’ or ‘require’),
but to actually use the 'xfce' package (exported by the ‘xfce’ module),
you need to add it to the ‘packages’ field of the ‘operating-system’.

items (packages, configuration files, etc) will be download or build
into the ’/gnu’ store when building the ’operating-system’, and packages
listed in the ‘packages’ field will be “installed” into the system
profile, aka ’/run/current-system/profile’.

so, the anwser is “Yes, Guix will install it automatically”.
Errors will be reported if a invalid configuration is used,
It’s a scheme file, so syntax and runtime errors may be common :-)