When modSha256 is null, disable the nix sandbox instead of using a
fixed-output derivation. This requires the nix-daemon to have
`sandbox = relaxed` set in their config to work properly.
Because the output is (hopefully) deterministic based on the inputs,
this should give a reproducible output. This is useful for development
outside of nixpkgs where re-generating the modSha256 on each mod.sum
changes is cumbersome.
Don't use this in nixpkgs! This is why null is not the default value.
https://hydra.nixos.org/build/108321871https://hydra.nixos.org/build/108556036
Things done:
* SSE2 isn't available for the aarch64-target of GCC (the package builds
on aarch64 without those flags and works fine as well).
* Added missing darwin libraries.
* Applied a darwin-specific patch which disables i386 as target platform
(breaks at least on MacOS catalina otherwise).
We made an effort to support ghcide in Nixpkgs, but the complexity of the
problem is a bit too high, IMHO. We need to keep older versions of several
packages around in order to satisfy the build requirements, and some of those
older packages don't even build themselves (like hie-bios). We had ghcide
working at some point, but then it was broken again right away after a couple
of days. I fear that we'll run into that issue again and again with a setup of
that complexity.
Instead, I'd propose that we work with upstream to fix their build, i.e. let's
make sure that the proper ghcide build works with recent versions of its build
inputs.
Closes https://github.com/NixOS/nixpkgs/pull/75449.
Closes https://github.com/NixOS/nixpkgs/pull/76103.
This PR fixes dhall_1_28_0, dhall-bash_1_0_25, and dhall-json_1_6_0 so
they build.
They all require a newer version of prettyprinter than we get from the
LTS package set.
This is from https://github.com/NixOS/nixpkgs/pull/75931 by @ijaketak.
Co-authored-by: Keito Kajitani <ijaketak@gmail.com>
