Seriously. Machine-readable semver is a pipe dream. You're relying on everyone else getting it right every single time. At best it's an indication of a maintain

Seriously. Machine-readable semver is a pipe dream. You're relying on everyone else getting it right every single time. At best it's an indication of a maintainer's intention. But it's not a substitute for QA; if your dependencies change, you need to test your software no matter what. So what does semver actually do? Using it as a guideline for setting your own version numbers to communicate to humans is one thing. Treating it as a machine-readable indicator with absolute certainty is taking it too far.