John Ericson

@japaric https://github.com/rust-lang/rfcs/pull/1133 while not completely solving the problems you mention is a big step in that direction.

We use this in NixOS too https://github.com/NixOS/nixpkgs/commit/694b4d29e1d0e4c86d497da437a3eed90910a7db it would be nice if it was upstreamed.

The only change this patch needs is to use `EXEEXT_FOR_BUILD` rather than `EXEEXT`.

If we could require `=`, i.e. `--foo=bar` rather than `--foo bar`, it would resolve the grammar ambiguity.

Do these flags effect building or just installing? I vaguely recall with `--enable-relocatable` its all `argv[0]`-relative, but I'm not sure otherwise. **edit** answer seems to be https://github.com/haskell/cabal/issues/3473

I don't think the old behavior is ever desirable. When cabal2nix is given source code it shouldn't be doing IO to get other source code, full stop. That just doesn't...

In https://github.com/haskell/cabal/issues/3882 @ezyang and I talked about turning Setup.hs into its own component? That probably dovetails with this nicely so setup components can be solved separately against the host compile...

@ezyang heh I actually missed the link to that patch reading through, I was just commenting on the issue in principle. Your plan there looks very good to me: "mixed...

Ok, to do this the Right Way™, I think I want to change the [type of qualifiers](https://github.com/haskell/cabal/blob/b32cfbcfce46f4232e9ff1a5cda3d895b9cd6ce7/cabal-install/Distribution/Solver/Types/PackagePath.hs#L43-L71). Note the comment on setup about nesting. Well, this applies equally well to...

@ezyang well in practice there would probably only be two, but I think the code is clearer using a list and supporting arbitrary.