Update nixpkgs unstable.

[shlevy]

Plutus needs NixOS/nixpkgs#127239.

[hamishmack]

bors try

[hamishmack]

bors try-

[hamishmack]

bors try

[iohk-bors]

try

Build failed:

• buildkite/haskell-dot-nix

[shlevy]

bors try

[iohk-bors]

🔒 Permission denied

Existing reviewers: click here to make shlevy a reviewer

[shlevy]

bors try

[iohk-bors]

🔒 Permission denied

Existing reviewers: click here to make shlevy a reviewer

[shlevy]

bors try

[shlevy]

Draft until NixOS/nixpkgs#133479 hits nixpkgs-unstable and hydra is happy

[iohk-bors]

try

Build failed:

• buildkite/haskell-dot-nix

[shlevy]

bors try

[iohk-bors]

try

Build failed:

• buildkite/haskell-dot-nix

[shlevy]

bors try

[iohk-bors]

try

Build failed:

• buildkite/haskell-dot-nix

[hamishmack]

bors try

[iohk-bors]

try

Build failed:

• buildkite/haskell-dot-nix

[hamishmack]

bors try

[iohk-bors]

try

Build failed:

• buildkite/haskell-dot-nix

[hamishmack]

bors try

[hamishmack]

bors try-

[hamishmack]

bors try

[iohk-bors]

try

Build failed:

• buildkite/haskell-dot-nix

[michaelpj]

I'm a bit confused about this PR: we can just bump nixpkgs in plutus, we pass in our own rather than using the haskell.nix pin.

[hamishmack]

bors try-

[hamishmack]

bors try

[hamishmack]

I'm a bit confused about this PR: we can just bump nixpkgs in plutus, we pass in our own rather than using the haskell.nix pin.

I was not sure how to do the equivalent of an ifdLevel 1 build on plutus and was worried it would build ghc at eval time. We made a PR to build it here first and it turned up some issues. First with ghc 8.10.5 on windows and once patched it seems we ghc 8.6.5 for windows is broken still. That would not be a problem if we were skipping the combination of nixpkgs-unstable, ghc 8.6.5 and windows. It turns out we have been including that version of GHC incorrectly via stack based tests that use a resolver with that version.

[shlevy]

@michaelpj We already bumped plutus, but would still be good to be able to move this forward as well. Also last I checked on this one of the failures was mingw cross, which may impact our plutus cross efforts.

[michaelpj]

I was not sure how to do the equivalent of an ifdLevel 1 build on plutus and was worried it would build ghc at eval time. We made a PR to build it here first and it turned up some issues

Ah yes, I remember now. It would be nice to figure out how to do that in downstream projects, but 👍

[iohk-bors]

try

Build succeeded:

• buildkite/haskell-dot-nix
• ci/hydra:Cardano:haskell-nix:required

[angerman]

Just to make this clear what @hamishmack said: nixpkgs-unstabe/ghc-8.6.5 are broken. This is due to the upated mingw in nixpkgs unstable. The required backport patch does not appear to be worth the effort to backport!