@Grimy Why do you think totient function would be possible in ECMA?

Well in ECMA 2018 it’s simply \B(?<!^\1*(?=\1*$)(xx+))

Returning via the number of matches is allowed by default: codegolf.meta.stackexchange.com/a/5338/6484

If you insist on returning via the length of the first match, it’s gonna need a much longer regex, but I see no obvious reason why it would be impossible.

I do insist on that. The number-of-matches thing is interesting, but cannot be used as a function in the context of a larger regex.

I mean, on that note, Fibonacci_N was also trivial.

Oh, overlapping matches

Afaik Perl6 is the only engine with an option to get all overlapping matches

Retina does it, so .NET might have that as a built-in mode.

@Grimy To be more specific: I would consider any such regex, returning the result as a number of matches, to be in a completely different language than those I would simply call "Regex (ECMAScript)" or similar.

For example if I added a switch to display number of matches in my engine, the name of the language would need to be "RegexMathEngine (with -c switch)" or something like that (with whatever I decided to name the switch).

Similarly to how I would have to submit regex solutions for non-decision-problem sequence questions (where you're asked to display the first N numbers or the Nth number in the sequence).

Anyway, I don't get why you're so confident the totient function is possible. I think it may well be impossible.

It seems that some trick like what I did with abundant numbers may be possible, but I wouldn't count on it.

It is however, currently the best candidate I can think of for something that might be impossible but might be possible. Another one I liked the idea of, but think is much more likely to be impossible, is π(x) (prime counting function).

And "is this an Ulam number?" seems very impossible.

Did I sound confident? I only said probably possible.

Agreed for Ulam numbers.