+1 for no one wants to deal with that regex. But I imagine plenty of people wondered how it would be done.

Thank you this code is very helpful specially on String manipulation.
I have learned that in the split method I can use regex on it and The \G

1. the code is very concise - that is good
2. readability is not very good due to the mysterious regexp
3. regarding productivity: I've tested a simple code execution in a loop (100000 times) and compared results with my implementation and it is almost 10 times slower (320 ms vs 43 ms) - and that is bad (most porbably either due to String.split complex implementation or because it compiles RegExp object on every call)

Cool solution an much better than my own, but in no way is that "best practice".

I wonder if regexes are considered "best practice" when there is a comparably easy to read for loop based answer?

Correct me if I am wrong: the String.split will find the pattern, remove it and keep left and right side.
So every 2 chars, split will remove pattern, keep left side (just got removed) and right side.
The trick here is that thanks to the capturing groups (?<=) + \G, pattern is reinserted(?) or pattern becomes "what's in between 2 chars" (?).
Still trying to figure it out.

That's a good idea.

how beautiful it is

Good ;)