Miss Dada  🏳️‍⚧️

Including Windows specific crates would be a much higher priority IMO. It's not that I'm worried that we aren't getting a sufficient sample, I just don't think we should pretend that it's more than exactly that: A sufficiently large sample

OK, I was concerned because you were talking about putting "too much trust" in a Crater run, but if you feel like it's by-and-large a sufficient sample (which I believe as well), then I think it's just a question of precise phrasing, which I don't care much about here.

I definitely don't think a crater run should be used as a justification for a breaking change, since not all Rust code is open source, and not all open source code is on http://crates.io . But for what it is trying to do, I think it's definitely a good sample

I mean, no question Crater is a used as a mean to gauge "permissible" breakage, i.e. things that are permitted by our official SemVer policy but where we don't want to cause significant pain in the ecosystem. Is that what you mean? If so, do you have an alternative in mind?

I wasn't aware that things like changing allowed kleene separators was under permissible breakage. Excluding std which has its supporting RFC, my understanding was that breaking changes were only permitted for bugs or soundness issues.

See https://github.com/rust-lang/rfcs/pull/1589 …, though I don't know if it applies here. But you said "frequently", so I assumed you were talking about library changes; I don't believe that language-level breaking changes have been frequent?

I'm not referring to bug fixes. I haven't really been keeping a list of things I've seen/been affected by (I've typically commented on things I've been affected by, but a lot of the time it's just easier to fix the breakage on my end than complain)