Unicode normalisation

x and y look the same:

x <- "å"
y <- stringi::stri_enc_nfd(x)
x
#> [1] "å"
y
#> [1] "å"

But are actually different:

nchar(x)
#> [1] 1
nchar(y)
#> [1] 2
x == y
#> [1] FALSE