I've been using semver for ages and I never realized it banned leading zeroes. Guess it makes sense, how many octal bugs were avoided?
Not sure how embarrassed to be by this, but when I don't recall knowing that when I first published calver.org. Despite having read the semver doc a dozen times at least...
Saw it recently and i am thinking of switching to it. It makes much, much more sense to instantly gauge maintainer activity in addition.
Easy on the integrator, easy on the maintainer, that's the idea. Especially with agents reading changelogs these days, time is the best human-legible heuristic.
Python rejected 3.YY versioning (PEP 2026) on a close vote. we were THIS close to Python versions meaning something.
A decade already, and the times have never been more timely. Apple, Nvidia, and many more on board, don't forget to check out the long-overdue timeline: https://calver.org/timeline.html
This is so fascinating I remember Apple's transition to versioning by year and thinking to myself "I wonder if anyone's standardized an approach to this in a more deterministic / calculated way"?
Cool to see someone has!
I'm heavily considering adopting this, would love to chat with anyone who has -- any tradeoffs here?