Pull bindings-uname package into Stack and fix it to work with FreeBSD.

[arrowd]

Due to #3515 I wasn't able to test if Stack builds with this change, but changes made for bindings-uname do work.

[ghost]

@mgsloan can you take a look at this PR ?

[mgsloan]

Looks like a reasonable change to me. I think it would be better to fix the bindings-uname package and update the one used by stack, though.

[arrowd]

I've never worked with Hackage. How can I update bindings-uname there?

[dbaynard]

It hasn't been updated since 2009, and there's no website, though the git repo still works. Ping the maintainer an email?

Edit: relevant haskell-infra/hackage-trustees#79

[arrowd]

Ping the maintainer an email?

Done that 8 days ago.

[dbaynard]

Yeah, I think this package needs a new maintainer. If you're willing to do it, there's a process on the hackage site.

The upshot of that is, I think we should merge the PR (subject to code review) and then when there's a new maintainer, migrate to the new code.

Which were the changes you made from the original bindings-uname?

[arrowd]

Which were the changes you made from the original bindings-uname?

Here is the patch: http://arrowd.name/uname.patch

[dbaynard]

Thanks. I'd ideally like to see that as two separate commits.

Due to #3515 I wasn't able to test if Stack builds with this change

Is this because there's no binary version of stack that works and so you can't bootstrap to test the changes? We should be getting a static build of stack again, soon — would that be helpful on that front?

It might be good to prioritize the fix to #3515 at the same time so it can be tested. That issue mentions the fix by @Blaisorblade for OpenBSD in 3cd92c8. Would it be possible to do a similar thing here?

[arrowd]

It has been tested by CI, it's all green. Doesn't that count?

[dbaynard]

It does count, although the integration tests aren't run by CI. I've asked maintainers to comment, but I think this will be OK to merge.

[borsboom]

I think it's OK to merge.

[arrowd]

I just found out that this PR also didn't make it into stable, so on this branch sysRelease still return garbage :<

I hope, this will be included in the next Stack release.