@kampka, thanks for your PR! By analyzing the history of the files in this pull request, we identified @coroa, @grahamc and @rycee to be potential reviewers.

Haven't had time to look at this but at some point I used the Python branch myself and this was the expression I used:

{ stdenv, fetchFromGitHub, python3Packages, xrandr }:

stdenv.mkDerivation rec {
  name = "autorandr-${version}";
  version = "20160509";

  src = fetchFromGitHub {
    owner = "phillipberndt";
    repo = "autorandr";
    rev = "6cd92b657027423721a29dfe0c7080483cdfc81e";
    sha256 = "01m8q8jqz8jmvsfpw1jb13yc11wrfim9hj7c496advp21krj7ynz";
  };

  buildInputs = [ python3Packages.python python3Packages.wrapPython ];

  patchPhase = ''
    substituteInPlace autorandr.py \
      --replace "popen(\"xrandr" "popen(\"${xrandr}/bin/xrandr" \
      --replace "\"xrandr\"" "\"${xrandr}/bin/xrandr\""
  '';

  installPhase = ''
    mkdir -p $out/bin $out/etc/bash_completion.d
    cp -v contrib/bash_completion/autorandr $out/etc/bash_completion.d
    cp -v autorandr.py $out/bin/autorandr

    wrapPythonPrograms
  '';

  meta = with stdenv.lib; {
    description = "Automatic display configuration selector based on connected devices";
    homepage = https://github.com/phillipberndt/autorandr/;
    maintainers = [ maintainers.coroa ];
    license = licenses.gpl3Plus;
  };
}

I don't know how relevant it is now but maybe there is something of interest to have there? I notice, for example, that the bash completion is not installed in this proposed PR.

Also, since the package is built from a git revision and not a formal release I believe the name should include "unstable" , i.e., name = "autorandr-unstable-${date}";. See https://nixos.org/nixpkgs/manual/#sec-package-naming

@rycee Thanks for sharing your expression, it was very useful.

I am aware of the naming policy, but I am unsure how renaming the derivation will impact the upgrade path.
Can you maybe shed some light on that?

@kampka if you mean the unstable prefix. you can change it without breaking anything.

@Mic92 Alright, done.

@kampka you have not uploaded any updates so far.

@Mic92 I have squashed the commits, sorry if that confuses the process.

Anyways, I noticed there is still an issue with the python version that I am having trouble fixing.
When using the batch mode, autorandr tryies to figure out its own binary [1], which fails if called from udev or systemd.
If someone has a good idea on how to patch / fix that, I'd be glad to hear it.

[1] https://github.com/phillipberndt/autorandr/blob/master/autorandr.py#L791

@kampka ok usually github rendered such changes differently.

@kampka if you mean the unstable prefix. you can change it without breaking anything.

Upgrading of imperatively installed programs (nix-env -i) uses the name of derivations to find the upgrade. So if I am correct add unstable might break that. We should document this behavior.

Anyways, I noticed there is still an issue with the python version that I am having trouble fixing.
When using the batch mode, autorandr tryies to figure out its own binary [1], which fails if called from udev or systemd.
If someone has a good idea on how to patch / fix that, I'd be glad to hear it.

Use substituteInplace ?

Ping?