Core-foundation-sys shouldn't have changed version at all

[tomaka]

Because its Cargo.toml contains a links attribute, it is impossible to use core-foundation-sys v0.2 and core-foundation-sys v0.3 in the same project.

There is absolutely no reason why the version of core-foundation-sys should match the version of core-foundation.

[tomaka]

It's too late now anyway, but it's a heads up for the next time.
I'm tired of fixing semver issues all the time. Now I have decided to report breakages to library owners.

[tomaka]

Hopefully the 24 people who watch this repo will be aware of this and will reject a PR that bumps core-foundation-sys.

[nox]

I am not sure why you employ such a tone to convey your annoyance, and this is not a "semver issue". While it seems this could have been 0.2.4, doing a breaking bump doesn't break semver.

I assume you are fed up because of rust-windowing/glutin#853, but this wouldn't have happened if glutin had said core-foundation = "0.2.3" and not core-foundation = "0". Just make sure to reject a PR that depends on crates with a "0" version constraint.

[tomaka]

@nox It has nothing to do with that but with the links attribute in the Cargo.toml.

[nox]

No, core-foundation 0.3 was selected by Cargo in glutin only because Cargo does weird stuff with a single digit "0" version. Had both winit and core-graphics said 0.2, none of them would have used 0.3 and no two different versions of core-foundation would have been used at the same time.

[tomaka]

I'm not talking about glutin's breakage here. It's totally glutin's fault and my fault for accepting a PR in the past that did = "0".

I'm opening this because we can't bump to core-foundation 0.3 without causing breakages in the entire ecosystem.