Inconsistent Versioning and the Case for Adopting SemVer

[monteiz]

Inconsistent Versioning and the Case for Adopting SemVer

[photonstorm]

This smells like an AI generated issue. We've covered how we handle versioning many times in the past (see our Dev Logs for details). Our Change Logs are extensive and detailed and always have been. The nicknames are never used to refer to a version in our documentation, it's all version number based. No, it doesn't strictly follow semver and at this stage in the projects life, it never can. That boat has sailed. When we move over to v4 versioning we can revisit it, but not until then.

[monteiz]

Yes, it is AI generated under my supervision. It's 10x faster and its English is surely better than mine :)

Happy to hear that SemVer is considered for v4.

I agree that the Change Logs are very extensive and I appreciate that. The benefit of SemVer is that you don't really need to read all the Change Logs every time (I think it's very unlikely that one is interested in all the changes, most likely a small part of it).

With SemVer, you just read the number and you know if your code will break or not, if you upgrade. That improves productivity a lot.

Plus, it forces you to strengthen your test suite to check retro-compatibility issues, which makes the framework more robust.