Bug in rm_repeated_characters() / ex_repeated_characters()

[sanjmeh]

library(qdapRegex)

"hha that was a good joke peep and pepper and pepe" |> rm_repeated_characters()
# "that was a good joke peep and pepper and"

"ahh that was a good joke peep and pepper and pepe"  |>  rm_repeated_characters()
#  "ahh that was a good joke peep and pepper and"

Note hha is removed while ahh is not removed. Is this by design or a bug? There is nothing in the documentation that says only starting characters will be checked.