You shouldn't spend your own effort; you should make it clear that you're open to users of such systems contributing.

That's how GCC became so dominant - there were people already using all sorts of Unixen and they wanted a compiler, so they made it work.

> You shouldn't spend your own effort; you should make it clear that you're open to users of such systems contributing.

In practice you can't really draw a line between those two things.

I don't know the end result, but I remember a discussion of how implementing parts of Git in Rust would be a problem because NonStop, a currently supported platform, has no Rust support. Of course the sane response would be "screw NonStop then", but having accepted contributions to make NonStop a supported platform that isn't an easy path to take.

That is absolutely true but exotic platforms are also fun to investigate and you can learn a lot. So I'd say you shouldn't spend your own effort if you don't want to but I am glad fractalfir did and I am looking forward to the his RustWeekNL presentation.

Of course, accepting contributions comes with some effort, too.

Funny, because the average person is convinced it's not worth spending any time supporting Linux!

I'm always convinced that people will pick up arbitrary projects that they are interested in and might not necessarily lead to a new pitch for venture capital or the next unicorn.