Ah, the Python package (python37Packages.monosat) build successfully too.

And maybe the Darwin/Linux closure size difference is because the Darwin build doesn't includeJava?

@GrahamcOfBorg build monosat python37Packages.monosat

Thanks! I'll try to test this over the weekend

Can we disable static builds and just have it shared? setup.sh already has it set to do --disable-static so that's a guideline for other packages.

I don't think that's straightforward to do. I think there would need to be changes made upstream to provide separate shared/static builds. I've made an issue on the monosat repo asking about this.

I've made a PR on the monosat repo to add an option to do a dynamically linked build only. If that gets merged I should be able to use that option for the Nix package. Would it not make sense to have that as a separate PR from this, since this is just fixing a currently broken build, whereas removing the static build is more of an improvement?

Your patch seems fine @acairncross. Can we use sambayless/monosat#24 downstream here?

Also, just for clarity, the build fails because it needs static libraries, which isn't really something we should curate in nixpkgs unless it's for the pkgsStatic overlay. So I'd think disabling them in anyway would be the most idiomatic fix to the build. Although in this case it required an improvement upstream.

My patch was merged upstream so I've bumped up the version here to use it. Is that ok?

The diff looks, umm... Huge? You should upgrade it to the beta release

• https://github.com/sambayless/monosat/releases/tag/1.8.0

have that in your commit message

monosat: $old_version -> $new_version

and apply that patch on top of it.

As you're the only (and new) maintainer of this I'm fine deferring to you as to what's acceptable.
Currently it's broken so it doesn't matter greatly.

Though I will mention, the preferred practice would be to make a derivative of the patch upstream to apply to this version... but as stated 😄

Edit: it appears the commit is in the 1.8.0 tag sambayless/monosat@033ac0f

The diff looks, umm... Huge?

Definitely a huge diff, although I think the majority of it is from applying a code formatter!

the preferred practice would be to make a derivative of the patch upstream to apply to this version

Ah, ok!

Anyway, I've bumped it to 1.8.0 (and done a bit of commit squashing).

Thanks @acairncross for the fix and the patience ✨ .
Luckily @Taneb brought this to my attention so we could get this merged.