doesn't report whether the build-from-source fallback worked or not, because that seems basically impossible
reports about how far each of the fetchers got (it does the fetch() -> bool for everything in parallel, and then tries the downloads sequentially. Knowing which ones were tried when they fail seems interesting to me.
only reports about the winning architecture on success (I think this is fine: we can kind-of infer the status of architectures higher in the preference list, and we don't really care about things lower down)
Also it turns out this is my first pull request to cargo-binstall. It took me a while to work out how things fit together. I added some comments in places where I had to dig down through too many layers of abstraction to find out the information that I was after.
similar to https://github.com/cargo-bins/cargo-binstall/pull/1905/files . I suspect that the discussions will be similar, but I won't push back on any changes you suggest.
Also it turns out this is my first pull request to cargo-binstall. It took me a while to work out how things fit together. I added some comments in places where I had to dig down through too many layers of abstraction to find out the information that I was after.