I like skeeto's version [0], but it allows for unbounded parsing due to unbounded zero padding of decimals, so therefore not constant time. Probably only requires an additional counter for the number of digits seen to mitigate that, but it would probably yucky the code a bit.

[0]: https://github.com/skeeto/scratch/blob/2af9b076117838b93fc2f...

Given that its 2025, a blog post that only covers parsing IPv4 addresses doesn't deserve much of my time.

I visited the blog post hoping the author would be presenting a dual-stack algorithm. I was sadly disappointed.