nixos/slim: reintroduce #75389

oxij · 2019-12-09T18:38:39Z

`git log`

Revert "Merge pull request nixos/slim: remove #73251 from worldofpeace/remove-slim"

This reverts commit 4583e29,
reversing changes made to d6fa540.

The fact that SLIM is not officially actively maintained is not a good
reason to remove it. It works, and, in some instances, like automatic
logins, it is the only thing that actually works. See
nixos: xserver.displayManager: use slim for automatic logins #46396 and
display-managers: make lightdm the default #30890 (comment).
slim: switch to slim-ng

... which is a fork maintained by me.
nixos/slim: save some closure space by using the default theme when using autologin

`nix-instantiate` environment

Host OS: Linux 4.9, SLNOS 19.09
Nix: nix-env (Nix) 2.2.2
Multi-user: yes
Sandbox: yes
NIXPKGS_CONFIG:

{
  checkMeta = true;
  doCheckByDefault = true;
}

Things done

Normal login works.
Autologin works.
Special commands work.

(Tested by applying this patchset on top of NixOS 19.09, and then testing the resulting buiild on a real hardware.)

rnhmjoj · 2019-12-09T19:59:47Z

nixos/doc/manual/release-notes/rl-2003.xml

@@ -148,13 +148,6 @@
      Now, all bash code is held to the same high standard, and the rather complex stateful manipulation of the options can be discarded.
    </para>
   </listitem>
-   <listitem>


Maybe mention instead that slim is now your slim-ng

Release notes should not be undone

jtojnar · 2019-12-09T23:07:11Z

The main issue with SLIM is that it does not integrate well with modern desktops like GNOME, Plasma or Sway. In particular:

It is missing the ability to load session files from $XDG_DATA_DIRS, or some other way of loading desktop files from more than a single sessiondir. While we are currently linking upstream session files into a single tree, $out/share/xsessions and $out/share/wayland-sessions will always be different directories (session type is determined from the directory name).
Wayland sessions are not supported.
Desktop environments are moving towards using systemd for starting sessions, and without upstream, adding support would fall on us.

It is misleading when a software with limited and frozen feature set is advertised as an alternative to modern maintained DMs. In order for SLIM to get back to tree, it should be de-emphasized in the documentation, and users should be warned about its state of development and supported features.

Though, to be honest, LightDM module is not on par with SDDM and GDM at the moment either (see e.g. #56342).

jtojnar · 2019-12-09T23:08:42Z

nixos/doc/manual/configuration/x-windows.xml

@@ -39,7 +39,7 @@
  can select an alternative one by picking one of the following lines:
 <programlisting>
 <xref linkend="opt-services.xserver.displayManager.sddm.enable"/> = true;
-<xref linkend="opt-services.xserver.displayManager.gdm.enable"/> = true;
+<xref linkend="opt-services.xserver.displayManager.slim.enable"/> = true;


Suggested change

<xref linkend="opt-services.xserver.displayManager.slim.enable"/> = true;

<xref linkend="opt-services.xserver.displayManager.gdm.enable"/> = true;

<xref linkend="opt-services.xserver.displayManager.slim.enable"/> = true; # limited integration

jtojnar · 2019-12-09T23:10:25Z

nixos/doc/manual/development/option-declarations.xml

@@ -146,7 +146,7 @@ xlink:href="https://nixos.org/nixpkgs/manual/#sec-package-naming">
      />), and to extend
   it in each backend module
   (<xref
-      linkend='ex-option-declaration-eot-backend-gdm' />,
+      linkend='ex-option-declaration-eot-backend-slim' />,


Suggested change

linkend='ex-option-declaration-eot-backend-slim' />,

linkend='ex-option-declaration-eot-backend-gdm' />,

<xref

linkend='ex-option-declaration-eot-backend-slim' />,

jtojnar · 2019-12-09T23:11:41Z

nixos/doc/manual/development/option-declarations.xml

@@ -167,11 +167,11 @@ services.xserver.displayManager.enable = mkOption {
 };</screen>
  </example>

-  <example xml:id='ex-option-declaration-eot-backend-gdm'>
-   <title>Extending <literal>services.xserver.displayManager.enable</literal> in the <literal>gdm</literal> module</title>
+  <example xml:id='ex-option-declaration-eot-backend-slim'>


Better use fully-featured DM here:

Suggested change

<example xml:id='ex-option-declaration-eot-backend-slim'>

<example xml:id='ex-option-declaration-eot-backend-gdm'>

jtojnar · 2019-12-09T23:12:24Z

nixos/doc/manual/development/option-declarations.xml

-  <example xml:id='ex-option-declaration-eot-backend-gdm'>
-   <title>Extending <literal>services.xserver.displayManager.enable</literal> in the <literal>gdm</literal> module</title>
+  <example xml:id='ex-option-declaration-eot-backend-slim'>
+   <title>Extending <literal>services.xserver.displayManager.enable</literal> in the <literal>slim</literal> module</title>


Suggested change

<title>Extending <literal>services.xserver.displayManager.enable</literal> in the <literal>slim</literal> module</title>

<title>Extending <literal>services.xserver.displayManager.enable</literal> in the <literal>gdm</literal> module</title>

jtojnar · 2019-12-09T23:12:55Z

nixos/doc/manual/development/option-declarations.xml

 <screen>
 services.xserver.displayManager.enable = mkOption {
-  type = with types; nullOr (enum [ "gdm" ]);
+  type = with types; nullOr (enum [ "slim" ]);


Suggested change

type = with types; nullOr (enum [ "slim" ]);

type = with types; nullOr (enum [ "gdm" ]);

jtojnar · 2019-12-09T23:17:37Z

pkgs/applications/display-managers/slim/default.nix

+  src = fetchFromGitHub {
+    owner = "oxij";
+    repo = pname;
+    rev = "b6d77d41eeb796d3788ecab2d02e616c858661e1";


Can we use a tag here?

jtojnar · 2019-12-09T23:19:20Z

pkgs/applications/display-managers/slim/default.nix

+
+  postPatch = ''
+    substituteInPlace CMakeLists.txt \
+      --replace /lib $out/lib


Aargh, why does not the build script use GNUInstallDirs CMake module?

The patch is ugly too. The build system should work in a really native way if it's literally a "For NixOS" package. GNUInstallDirs would be the way to go for that.
(though, not very fond of CMake)

jtojnar · 2019-12-09T23:19:53Z

pkgs/applications/display-managers/slim/default.nix

+
+  cmakeFlags = [ "-DUSE_PAM=1" ];
+
+  enableParallelBuilding = true;


CMake builds in parallel by default.

jtojnar · 2019-12-09T23:21:14Z

pkgs/applications/display-managers/slim/default.nix

+  enableParallelBuilding = true;
+
+  buildInputs =
+    [ cmake pkgconfig libjpeg libpng fontconfig freetype


Build platform dependencies like cmake and pkg-config should go to nativeBuildInputs

pkgs/applications/display-managers/slim/default.nix

oxij · 2019-12-10T03:38:05Z

@rnhmjoj

Mybe mention instead that `slim` is now your `slim-ng`

What for? It is a drop-in replacement. @jtojnar

* It is missing the ability to load session files from `$XDG_DATA_DIRS`, or some other way of loading desktop files from more than a single `sessiondir`. While we are currently linking upstream session files into a single tree, `$out/share/xsessions` and `$out/share/wayland-sessions` will always be different directories (session type is determined from the directory name).

This would very easy to implement, assuming this is ever required. But since Wayland support is unlikely (because pretty much everything will need to be rewritten) I don't really see a problem.

* Wayland sessions are not supported.

Yes. Wayland support is unlikely. Unless there exists a small library that would allow to reimplement the rendering code for both X11 and Wayland in a single breath. (Hm. Sounds like SDL. Rewriting it for SDL sounds doable and maybe even enjoyable, so maybe. Not soon, thought, I have other plans for quite awhile.)

* Desktop environments are moving towards using [systemd for starting sessions](https://blogs.gnome.org/laney/2018/06/26/starting-sessions-with-systemd/), and without upstream, adding support would fall on us.

Looks fairly easy to implement, thought I don't particularly like the idea. IMHO, if you want to run stuff with systemd-user you can just plug it into a *.desktop file or whatever. Why should DM even care about it? Sound like another not-invented-here-so-systemd-will-have-reinvent-this-one-too thing to me.

It is misleading when a software with limited and frozen feature set is advertised as an alternative to modern maintained DMs. In order for SLIM to get back to tree, it should be de-emphasized in the documentation, and users should be warned about its state of development and supported features.

Sure. Will fix.

avnik · 2019-12-10T16:55:31Z

@jtojnar I am sure, that most of slim (or slim-ng) users don't use major DEs, and use it with simple WMs (icewm/xmonad/qtile/etc). It can be mentioned in docs as recomendation to choose other one for GNOME, etc.

worldofpeace · 2019-12-11T05:27:02Z

@oxij This package shouldn't even be called slim. It isn't slim anymore, it's a package you have taken ownership of and continued. It's likely to proliferate in use. You should even rename the slim module to slim-ng.

Also, what about:

systemd services/session
the bug in the screen locker Slimlock Bug/Vulnerability: isn't guaranteed to grab keyboard iwamatsu/slim#4
issues mentioned by @jtojnar

I'd also want to emphasize the importance of xdg spec compliance.

And the biggest issue is wayland, it can't start wayland sessions.
You don't actually have to rewrite it to run on wayland, at minimum could it start a wayland-session?
We did mention that lightdm doesn't have this capability, but it's only on NixOS.
I've recently contributed canonical/lightdm@03f2189, and it should soon work properly on nixos.

So I'm seeing this being blocked on some code changes.
I'm not equipped to review C++ code, since you are making changes to slim's code I feel the need for someone with the know-how to review your work and diff the changes from the prior project.

I mentioned how if a community member forked and continued the project we'd be in good hands.
I'm glad this has happened.

worldofpeace · 2019-12-11T05:28:41Z

nixos/tests/slim.nix

+    startAll;
+    $machine->waitForText(qr/Username:/);
+    $machine->sendChars("${user.name}\n");
+    $machine->waitForText(qr/Password:/);
+    $machine->sendChars("${user.password}\n");
+
+    $machine->waitForFile('${user.home}/.Xauthority');
+    $machine->succeed('xauth merge ${user.home}/.Xauthority');
+    $machine->waitForWindow('^IceWM ');
+
+    # Make sure SLiM doesn't create a log file
+    $machine->fail('test -e /var/log/slim.log');
+  '';


Can we port this to the python driver? #72828

This should just be replacing make-test.nix with make-test-python.nix above and using:

start_all() machine.wait_for_text("Username:") machine.send_chars("${user.name}\n") machine.wait_for_text("Password:") machine.send_chars("${user.password}\n") machine.wait_for_file("${user.home}/.Xauthority") machine.succeed("xauth merge ${user.home}/.Xauthority") machine.wait_for_window(r"^IceWM ") # Make sure SLiM doesn't create a log file machine.fail("test -e /var/log/slim.log")

worldofpeace · 2019-12-11T05:29:14Z

pkgs/applications/display-managers/slim/default.nix

+  buildInputs =
+    [ cmake pkgconfig libjpeg libpng fontconfig freetype
+      pam dbus
+      xorg.libX11 xorg.libXext xorg.libXrandr xorg.libXrender xorg.libXmu xorg.libXft makeWrapper


what has makeWrapper been used for?

if makeWrapper is used it should be in nativeBuildInputs as well.

worldofpeace · 2019-12-11T05:34:15Z

Can slim-ng support default or soon to be defaultSession?

rnhmjoj · 2019-12-11T09:07:46Z

What for? It is a drop-in replacement.

Technically It's a different software. One should know it's getting your fork and not the original slim.

This package shouldn't even be called slim. It isn't slim anymore, it's a package you have taken ownership of and continued. It's likely to proliferate in use. You should even rename the slim module to slim-ng.

Is there a policy on this topic? For example in 34cc444 compton has been (more or less) silently replaced by a fork. Imho even if by the same author this should be mentioned in the release notes and maybe the attribute name should change as well.

worldofpeace · 2019-12-11T10:13:14Z

Is there a policy on this topic? For example in 34cc444 compton has been (more or less) silently replaced by a fork. Imho even if by the same author this should be mentioned in the release notes and maybe the attribute name should change as well.

In that particular example, I don't think it should even have been aliased. An alias could be used when two identities were created for the same thing, old and new. A good example was gnome-mpv was renamed to celluloid, but it's the same code. In that example you want to help out with migraton, so you create a throw indicating that compton has been removed, but picom is essentially a drop in replacement. This is what should be done in slim-ng as well.

worldofpeace · 2019-12-11T10:18:27Z

However, sometimes in the cases like with forks, the new maintainer doesn't re-identify this software. Things become historically confusing, and the "new" source will proliferate in distributions but it's fundamentally different. It's not as bad because it's usually 1:1 compatible, but as a packager, I'd request the maintainer to re-identify the software to avoid this.

So to make on topic with slim, we shouldn't call this slim because we're not talking about the project without a developer who also proclaimed it abandoned, and with various compatibility issues.

oxij · 2019-12-13T12:05:29Z

> @@ -39,7 +39,7 @@ can select an alternative one by picking one of the following lines: <programlisting> <xref linkend="opt-services.xserver.displayManager.sddm.enable"/> = true; -<xref linkend="opt-services.xserver.displayManager.gdm.enable"/> = true; +<xref linkend="opt-services.xserver.displayManager.slim.enable"/> = true; ```suggestion <xref linkend="opt-services.xserver.displayManager.gdm.enable"/> = true; <xref linkend="opt-services.xserver.displayManager.slim.enable"/> = true; # limited integration ```

Done.

> @@ -146,7 +146,7 @@ xlink:href="https://nixos.org/nixpkgs/manual/#sec-package-naming"> />), and to extend it in each backend module (<xref - linkend='ex-option-declaration-eot-backend-gdm' />, + linkend='ex-option-declaration-eot-backend-slim' />, ```suggestion linkend='ex-option-declaration-eot-backend-gdm' />, <xref linkend='ex-option-declaration-eot-backend-slim' />, ``` > @@ -167,11 +167,11 @@ services.xserver.displayManager.enable = mkOption { };</screen> </example> - <example xml:id='ex-option-declaration-eot-backend-gdm'> - <title>Extending <literal>services.xserver.displayManager.enable</literal> in the <literal>gdm</literal> module</title> + <example xml:id='ex-option-declaration-eot-backend-slim'> Better use fully-featured DM here: ```suggestion <example xml:id='ex-option-declaration-eot-backend-gdm'> ``` > @@ -167,11 +167,11 @@ services.xserver.displayManager.enable = mkOption { };</screen> </example> - <example xml:id='ex-option-declaration-eot-backend-gdm'> - <title>Extending <literal>services.xserver.displayManager.enable</literal> in the <literal>gdm</literal> module</title> + <example xml:id='ex-option-declaration-eot-backend-slim'> + <title>Extending <literal>services.xserver.displayManager.enable</literal> in the <literal>slim</literal> module</title> ```suggestion <title>Extending <literal>services.xserver.displayManager.enable</literal> in the <literal>gdm</literal> module</title> ``` > <screen> services.xserver.displayManager.enable = mkOption { - type = with types; nullOr (enum [ "gdm" ]); + type = with types; nullOr (enum [ "slim" ]); ```suggestion type = with types; nullOr (enum [ "gdm" ]); ```

Basically, you want me to edit examples in `doc/manual/development/option-declarations.xml`. I think this a) should be outside of the scope of this PR, b) makes no sense, these are examples for module developers, not `configuration.nix` examples.

> @@ -261,6 +261,7 @@ in rec { inherit require; virtualisation.memorySize = 1024; services.xserver.enable = true; + services.xserver.displayManager.slim.enable = false; IMHO, we should keep the default DM in tests. > @@ -248,6 +248,7 @@ in rec { inherit require; virtualisation.memorySize = 1024; services.xserver.enable = true; + services.xserver.displayManager.slim.enable = false; Same here.

Done.

> @@ -1,9 +1,9 @@ # This module declares the options to define a *display manager*, the -# program responsible for handling X logins (such as LightDM, GDM, or SDDM). -# The display manager allows the user to select a *session -# type*. When the user logs in, the display manager starts the +# program responsible for handling X logins (such as xdm, gdb, or ```suggestion # program responsible for handling X logins (such as LightDM, GDM, SDDM, or ```

Done.

> @@ -321,14 +321,15 @@ in execCmd = mkOption { type = types.str; example = literalExample '' - "''${pkgs.lightdm}/bin/lightdm" + "''${pkgs.slim}/bin/slim" Let’s keep it on the down low. ```suggestion "''${pkgs.lightdm}/bin/lightdm" ``` > ''; description = "Command to start the display manager."; }; environment = mkOption { type = types.attrsOf types.unspecified; default = {}; + example = { SLIM_CFGFILE = "/etc/slim.conf"; }; I do not think it is good idea to include it in the example, as it will conflict with the SLiM module when used.

Done.

> @@ -0,0 +1,42 @@ +{ stdenv, fetchFromGitHub, fetchpatch, cmake, pkgconfig, xorg, libjpeg, libpng

All of those done.

oxij · 2019-12-13T12:23:59Z

* the bug in the screen locker iwamatsu/slim#4

I think this should be closed by 48d25b6969c1f36fe6b505f7eb308cb1771d9d2b of slim-ng, but at the moment I wouldn't recommend using slimlock anyway, it's very far from what a proper screen locker (like physlock) should be doing. IMHO, the code there needs a good reorganization before those issues can be fixed in a sane manner (e.g., currently slimlock copy-pastes a bunch of code from app.cpp, uhg!).

* issues mentioned by @jtojnar

Those are fixed.

Can we port this to the python driver? #72828

Out of scope of this PR.

Can slim-ng support `default` or soon to be [`sessionDefault`](#53843)?

That needs more research. The cursory look didn't help. Hopefully, this is not blocking?

slim -> slim-ng

I'm not sure what you want here. Renaming the nixos module seems rather strange. Even if slim-ng would diverge wildly from slim I would rather keep the same module and add an option there to select between packages, since they would do essentially the same thing. Moreover, since, at this moment, slim-ng is just slim with less bugs I would vote against renaming the package attribute too. Why add an alias when you can skip it?

core review

Sure, if somebody wants to do it, feel free. But, in general, the code there is rather messy so don't blame me for something I didn't do, please. With slim-ng v1.3.8 I did my best without, hopefully, breaking anything.

worldofpeace · 2019-12-17T06:05:16Z

Out of scope of this PR.

Porting to the python testing driver is not actually of of scope.
Because this is reintroducing a test, all tests need to be written in python as the perl driver is due for removal and tests are actively being moved over #72828.
I have no reason to treat this test like any other test being added.

worldofpeace · 2019-12-17T06:09:27Z

I'm not sure what you want here. Renaming the nixos module seems rather strange. Even if slim-ng would diverge wildly from slim I would rather keep the same module and add an option there to select between packages, since they would do essentially the same thing.

Moreover, since, at this moment, slim-ng is just slim with less bugs I would vote against renaming the package attribute too. Why add an alias when you can skip it?

Looking at some commits it's slowly diverting from slim, and already behaves in slightly different ways.
In my personal experience, I think a rename could be a good move here.

core review
Sure, if somebody wants to do it, feel free. But, in general, the code there is rather messy so don't blame me for something I didn't do, please. With slim-ng v1.3.8 I did my best without, hopefully, breaking anything.

I will relay it to the person that any design decisions or original code is not to come back and reflect on you directly 😄

worldofpeace · 2019-12-21T00:47:53Z

The main issue with SLIM is that it does not integrate well with modern desktops like GNOME, Plasma or Sway.

It is amazing with XMonad.

The desktops mentioned there provide wayland sessions. As X has no future this should be a great concern.

rnhmjoj · 2020-01-16T20:42:21Z

Updates? Also, is the problem in #4300 still present?

misuzu · 2020-01-30T16:06:34Z

slim seems to work fine, but slimlock theme is taken neither from configuration nor from SLIM_THEMESDIR variable.

misuzu · 2020-02-02T10:52:01Z

I think slimlock has always been un-configurable because lack of slimlock.conf.

slimlock from regular slim do take theme from SLIM_THEMESDIR env variable so this is kind of regression.

This reverts commit 4583e29, reversing changes made to d6fa540. The fact that SLiM is not officially actively maintained is not a good reason to remove it. It works, and, in some instances, like automatic logins, it is the only thing that actually works. See NixOS#46396 and NixOS#30890 (comment).

FRidh

Given there is apparently quite the interest, I suppose with the users taking up active maintainership it should be fine to get this back in. Regarding compatibility, is it possible to add a check preventing it from being used with incompatible DE's?

FRidh · 2020-02-10T19:48:05Z

nixos/doc/manual/release-notes/rl-2003.xml

@@ -148,13 +148,6 @@
      Now, all bash code is held to the same high standard, and the rather complex stateful manipulation of the options can be discarded.
    </para>
   </listitem>
-   <listitem>


Release notes should not be undone

FRidh · 2020-02-10T19:49:02Z

nixos/tests/slim.nix

+  name = "slim";
+
+  meta = with pkgs.stdenv.lib.maintainers; {
+    maintainers = [ aszlig ];


FRidh · 2020-02-10T19:49:19Z

nixos/modules/services/x11/display-managers/slim.nix

+    security.pam.services.slimlock = {};
+
+    environment.systemPackages = [ pkgs.slim ];
+


add as maintainer

…sing autologin

oxij · 2020-02-10T21:28:36Z

@jtojnar

tests

Done. No idea if it works, though. @rnhmjoj

Also, is the problem in #4300 still present?

Hm, I use it pretty much the same way as the author there, but I see no such thing here. I'll check carefully later. @misuzu @bb2020

slimlock

Should be fixed now.

Also xserver.displayManager.setupCommands was not available in the previous version. Maybe we can find a way to implement it now.

Yes, we can implement it now. Will do later. @FRidh

Regarding compatibility, is it possible to add a check preventing it from being used with incompatible DE's?

Sure, but I don't know which ones are incompatible, I only use xmonad.

Release notes should not be undone

Well, what should I do then? That release note is no longer valid. Should I add another one there saying that it was re-added again or what? @ ALL Also renamed the attribute to `slim-ng` and added an alias on popular demand.

misuzu · 2020-02-14T11:30:44Z

slimlock themes is working now, but there are still some issues. Try doing this:

boot your machine with slim enabled
enter your username and wrong password
enter your username and right password
observe crash without proper restart

worldofpeace · 2020-02-14T20:03:28Z

pkgs/top-level/aliases.nix

  smbclient = samba; # added 2018-04-25
-  slim = throw "slim has been removed. Please use a different display-manager"; # added 2019-11-11
-  slimThemes = throw "slimThemes has been removed because slim has been also"; # added 2019-11-11
+  slim = slim-ng; # added 2019-11-11
  net_snmp = net-snmp; # added 2019-12-21


This isn't correct. An alias is supposed to be a bind to something that's equivalent, and we've already addressed that it isn't anymore.

slim = throw "slim has been removed. Please use the replacement slim-ng or a different display-manager"

wmertens

I agree that there is a place for slim-ng in NixOS, but IMHO it should be called slim-ng and clearly mention the differences with gdm/sddm in the description of the module

wmertens · 2020-06-04T21:08:45Z

nixos/doc/manual/configuration/x-windows.xml

@@ -39,7 +39,7 @@
  can select an alternative one by picking one of the following lines:
 <programlisting>
 <xref linkend="opt-services.xserver.displayManager.sddm.enable"/> = true;
-<xref linkend="opt-services.xserver.displayManager.gdm.enable"/> = true;


Why remove gdm? That's what most users probably want

wmertens · 2020-06-04T21:10:22Z

nixos/doc/manual/development/option-declarations.xml

@@ -99,7 +99,7 @@ xlink:href="https://nixos.org/nixpkgs/manual/#sec-package-naming">
  <para>
   As an example, we will take the case of display managers. There is a central
   display manager module for generic display manager options and a module file
-   per display manager backend (sddm, gdm ...).
+   per display manager backend (slim, sddm, gdm ...).
  </para>


Putting slim first makes it seem like the first one users should try, whereas slim is a very leftfield alternative

stale · 2021-06-04T23:56:01Z

I marked this as stale due to inactivity. → More info

oxij requested a review from nbp as a code owner December 9, 2019 18:38

oxij mentioned this pull request Dec 9, 2019

nixos/slim: remove #73251

Merged

10 tasks

rnhmjoj reviewed Dec 9, 2019

View reviewed changes

jtojnar reviewed Dec 9, 2019

View reviewed changes

worldofpeace reviewed Dec 11, 2019

View reviewed changes

worldofpeace added the 2.status: work-in-progress This PR isn't done label Dec 11, 2019

oxij force-pushed the pkgs/readd-slim branch from ee64265 to b650ba6 Compare December 13, 2019 12:32

FRidh reviewed Feb 10, 2020

View reviewed changes

jtojnar and others added 4 commits February 10, 2020 21:13

nixos/tests/slim: migrate to python

debc8b2

slim: switch to slim-ng

60c8c36

nixos/slim: deemphasize

8a54abc

nixos/slim: save some closure space by using the default theme when u…

d0d47d0

…sing autologin

oxij force-pushed the pkgs/readd-slim branch from b650ba6 to d0d47d0 Compare February 10, 2020 21:25

worldofpeace reviewed Feb 14, 2020

View reviewed changes

wmertens requested changes Jun 4, 2020

View reviewed changes

misuzu mentioned this pull request Jul 24, 2020

nixos/displayManager: slimlock fix (#67350) #67751

Closed

10 tasks

ryantm added 2.status: merge conflict This PR has merge conflicts with the target branch and removed 2.status: merge conflict This PR has merge conflicts with the target branch labels Oct 3, 2020

stale bot added the 2.status: stale https://github.com/NixOS/nixpkgs/blob/master/.github/STALE-BOT.md label Jun 4, 2021

Artturin closed this May 7, 2022

oxij mentioned this pull request Jul 17, 2023

Typed nixpkgs.config with Gentoo-like use-flags #56227

Closed

	<xref linkend="opt-services.xserver.displayManager.slim.enable"/> = true;
	<xref linkend="opt-services.xserver.displayManager.gdm.enable"/> = true;
	<xref linkend="opt-services.xserver.displayManager.slim.enable"/> = true; # limited integration

	<example xml:id='ex-option-declaration-eot-backend-slim'>
	<example xml:id='ex-option-declaration-eot-backend-gdm'>

	<title>Extending <literal>services.xserver.displayManager.enable</literal> in the <literal>slim</literal> module</title>
	<title>Extending <literal>services.xserver.displayManager.enable</literal> in the <literal>gdm</literal> module</title>

	type = with types; nullOr (enum [ "slim" ]);
	type = with types; nullOr (enum [ "gdm" ]);


		cmakeFlags = [ "-DUSE_PAM=1" ];

		enableParallelBuilding = true;

		security.pam.services.slimlock = {};

		environment.systemPackages = [ pkgs.slim ];

nixos/slim: reintroduce #75389

nixos/slim: reintroduce #75389

Conversation

oxij commented Dec 9, 2019

git log

nix-instantiate environment

Things done

rnhmjoj Dec 9, 2019 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jtojnar commented Dec 9, 2019

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

oxij commented Dec 10, 2019 via email

avnik commented Dec 10, 2019

worldofpeace commented Dec 11, 2019 • edited Loading

Choose a reason for hiding this comment

jtojnar Dec 21, 2019 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

worldofpeace commented Dec 11, 2019 • edited Loading

rnhmjoj commented Dec 11, 2019

worldofpeace commented Dec 11, 2019

worldofpeace commented Dec 11, 2019 • edited Loading

oxij commented Dec 13, 2019 via email

oxij commented Dec 13, 2019 via email

worldofpeace commented Dec 17, 2019

worldofpeace commented Dec 17, 2019

worldofpeace commented Dec 21, 2019

rnhmjoj commented Jan 16, 2020

misuzu commented Jan 30, 2020

misuzu commented Feb 2, 2020

FRidh left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

oxij commented Feb 10, 2020 via email

misuzu commented Feb 14, 2020

Choose a reason for hiding this comment

wmertens left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

stale bot commented Jun 4, 2021

`git log`

`nix-instantiate` environment

rnhmjoj Dec 9, 2019 •

edited

Loading

worldofpeace commented Dec 11, 2019 •

edited

Loading

jtojnar Dec 21, 2019 •

edited

Loading

worldofpeace commented Dec 11, 2019 •

edited

Loading

worldofpeace commented Dec 11, 2019 •

edited

Loading