[ryukoposting]

Conventional commits is one of the weirder cargo-culty things we do in this industry. I do take issue with this, though:

> Automatically determining a semantic version bump (based on the types of commits landed): This sounds nice, but the realities of software engineering often interfere significantly with the viability of accurately accomplishing this... imagine a situation where the breaking change you introduced was actually so breaking that you have to revert it... maybe the breakage is subtle and you don’t realise a change is a breaking change when you make the change. Only in retrospect realise that it’s breaking. You will incorrectly increment a minor/patch version when a major version bump is necessary...say you later add a commit which, in composition with a previously breaking commit, results in a diff which is not breaking. Similar to the revert situation, tooling would incorrectly identify a breaking change.

This is a problem with Semver, not conventional commits. You're liable to do this regardless of how your commits are formatted, which is one of several reasons why conventional commits are silly. But in this case, Semver itself drew semantic lines that aren't clear, and are easily broken.