[haskell/generic-builder] windows always has an active library

[angerman]

Motivation for this change

Broken out of #43559

Things done

• Tested using sandboxing (nix.useSandbox on NixOS, or option sandbox in nix.conf on non-NixOS)
• Built on platform(s)
  • NixOS
  • macOS
  • other Linux distributions
• Tested via one or more NixOS test(s) if existing and applicable for the change (look inside nixos/tests)
• Tested compilation of all pkgs that depend on this change using nix-shell -p nox --run "nox-review wip"
• Tested execution of all binary files (usually in ./result/bin/)
• Determined the impact on package closure size (by running nix path-info -S before and after)
• Fits CONTRIBUTING.md.

---

[peti]

I am sorry, but I don't understand the reason behind this change. I also read the commit message and now I am even more confused than before. Could you please re-state your reasons for making this change to help me understand what you mean by "the check is wrong"? And why is Windows different than other platforms?

[angerman]

@peti, fair point, this is a rather confusing PR. And I'm not even happy with it.

Here's the underlying issue: nixpkgs assumes that there is a library if it's either dynamic, or static, or profiling; and it's a library to begin with. This however is not correct. You can have a library that is neither --enable-static, nor --enable-shared. That is you end up with a library that is just a collection of obejct files as well as an aggregated archive.

The underlying issue is that --enable-static with cabal passes -staticlib to ghc, and -staticlib does not work on windows, and as such you can't set enableStaticLibraries, but at the same time if you don't want enableSharedLibraries, you end up with empty libraries, which is certainly not what you want.

[peti]

I'll just trust your judgement going forward with this. Please do whatever you think is best. :-)

[matthewbauer]

I think this is an okay thing for now. Especially if it gets us closer to ghc/mingw support.

[angerman]

To be fair, I think the correct solution is to get rid of the hasActiveLibrary logic, and just go with isLibrary.
The reasoning is as follows:

• enableStaticLibraries -> hasActiveLibrary
• enableSharedLibraries -> hasActiveLibrary
• enableLibraryProfiling -> hasActiveLibrary
  and as I outlined above, !enableStaticLibraries && !enableSharedLibraries && !enableLibraryProfiling should not imply !hasActiveLibrary.

[domenkozar]

target haskell-updates branch

[angerman]

@domenkozar alright :-) haskell-updates it is.

[ryantm]

I didn't look at this carefully, but I'm going to merge it based on all the approval it has. Let's see how it does on Hydra!