I think the core issue is that wc is not a super optimized program. It is sufficiently fast for most purposes and so hasn’t ever been improved.
https://github.com/ojhunt/wc
It's the simplest thing that could obviously work, but its generally not tested beyond the most basic versions.
That said I’ll try to post my horrifying impl to GitHub later :)
I think the core issue is that wc is not a super optimized program. It is sufficiently fast for most purposes and so hasn’t ever been improved.