Use SemVer properly for breaking changes

[pwbriggs]

I was delighted to see a new release of MA this afternoon. However, I was not so happy to see that, despite 2.8.0 being designated a minor release under SemVer, there was a big warning at the top of the release notes (I'm glad I glanced at them!) saying that this release does, in fact, include breaking changes. This undermines the purpose of using SemVer, in my opinion—I believe that any breaking change should increment the major counter so that users can more effectively gauge the amount of research they should do before upgrading. Contrary to how some projects operate, a "major release" shouldn't have to mean major changes were made—simply breaking changes. Check out Tom Preston-Werner's "Major Version Numbers are Not Sacred" for a much more qualified argument about why breaking changes should always be major releases. Thanks!

[marcelveldt]

It is not really that big of a breaking change. We noted it that way to draw attention to it. For 99% of the people this won't be a breaking change.

So no, I'm not agreeing this should be a 3.0 version. We are not blindly following SemVer at all.