Hacker News new | comments | show | ask | jobs | submit login

I think it is unlikely we will make FreeBSD tier 1 any time soon. The project is stretched beyond its limits with its platform support commitments, and FreeBSD is historically difficult for us to support.

The next tier 1 is going to be ARM, probably for both Android and Linux, and then I personally don't see any others on the horizon. I actually see our tiers being redefined soon to be more graduated, and not have so many tier 2 commitments (we're happy to build artifacts for obscure platforms, less happy to be on the hook for guaranteeing obscure platforms continue to build - not saying FreeBSD specifically is 'obscure').

That said, the differences between tier 1 and tier 2 are not so great. If a platform has testing on CI then it is pretty close to tier 1, even if it's technically tier 2. The problem here is that testing the FreeBSD targets requires running FreeBSD in some way and that has been hard for our CI to do. Anything that can be crossed from Linux is easy, everything else is a painful special case. So with FreeBSD we do the the cross-building, but not the cross-testing. So the most effective thing to do to take Rust FreeBSD support to the next level is to figure out a way to get the CI to run tests.

There may be opportunities in the future for paid Rust platform support, where if a motivated party donates the right amount of money, we'll make it happen. I think that will have to happen to get some of the remaining platforms up to production quality.




>So the most effective thing to do to take Rust FreeBSD support to the next level is to figure out a way to get the CI to run tests.

If I were to set up my own FreeBSD buildbot for commits to the Rust repos, would result reports from my buildbot be welcome? Would automatic reports be useful to you, or should I manually inspect failures and write up a bug report / pull request with fixes for issues that arise? If automatic, how could I best integrate reports of failure with your development process for Rust?

>There may be opportunities in the future for paid Rust platform support, where if a motivated party donates the right amount of money, we'll make it happen. I think that will have to happen to get some of the remaining platforms up to production quality.

I like this idea though I don't have a lot of money to spend myself currently.


We no longer use buildbot for our CI.

We have a "meta CI" system called bors/homu that integrates with other CI, currently Travis and AppVeyor. bors works with any service that can post a result to a GitHub PR.

So the plan for supporting further architectures is to create a small custom runner that builds, tests and posts to the PR. That code does not exist yet, but if it appeared there are likely several upcoming architectures that could make use of it.

With that tool, we could allow contributors to set up their own integration bots and hook into our CI.


I've got scripts that I use for my own kernel development that spin up a VM (using kvm), run a test inside the VM and give you pass/fail and test output on stdout.

If anyone is interested, it'd probably be quite easy to adapt for running rust tests inside a FreeBSD vm:

https://evilpiepirate.org/git/ktest.git

If anyone is interested I'd be happy to walk someone through what it does and what parts you'd want.




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: