I’m new to Nix, and trying to formulate a flake for a package where I’m part of the development team.
The general package seems to work, but I now want to add some of the optional dependencies. One of these dependencies is the cbc package, which has an old codebase. Because it uses the register type, the package definition contains the following line: https://github.com/NixOS/nixpkgs/blob/d6a0fea30ee3921729a2d62611cd7f097f4d0409/pkgs/applications/science/math/cbc/default.nix#L17
This seems to work correctly, because when I run nix shell nixpkgs#cbc I get an environment with a working cbc executable.
Building my flake, with the definition underneath, without the cbc dependency works. However, trying in my flake with the cbc package dependency, the nix build . command will fail with many errors like:
CoinOslFactorization3.cpp:2594:15: error: ISO C++17 does not allow 'register' storage class specifier
It seems to me like the rule enforcing C++14 is not taken into account. Am I doing something wrong? How can I resolve this problem?
flake.nix
{
description = "Medium-level constraint modelling language";
inputs = {
nixpkgs.url = "github:NixOS/nixpkgs/nixos-unstable";
flake-utils.url = "github:numtide/flake-utils";
};
outputs = { self, nixpkgs, flake-utils }:
flake-utils.lib.eachDefaultSystem (system:
let
pkgs = nixpkgs.legacyPackages.${system};
in
{
packages = {
minizinc = pkgs.callPackage ./default.nix { };
default = self.packages.${system}.minizinc;
};
devShells.default = pkgs.mkShell {
name = "minizinc";
inputsFrom = [ self.packages.${system}.minizinc ];
packages = with pkgs; [
bison
flex
ccache
];
};
});
}
default.nix
{ lib, stdenv, cbc, cmake, gecode, mpfr, ninja, zlib }:
stdenv.mkDerivation {
name = "minizinc";
src = ./.;
nativeBuildInputs = [ cmake ninja ];
buildInputs = [ gecode mpfr zlib cbc ];
meta = with lib; {
homepage = "https://www.minizinc.org/";
description = "A medium-level constraint modelling language";
longDescription = ''
MiniZinc is a medium-level constraint modelling
language. It is high-level enough to express most
constraint problems easily, but low-level enough
that it can be mapped onto existing solvers easily and consistently.
'';
license = licenses.mpl20;
platforms = platforms.unix;
};
}