CH-LanguagesCharacterSets.xml
/div/div[2]/div[7]/div[2]/p[1]
after
seems to be most appropriate for text encoding projects.
add
NFC as far as possible converts all base characters followed by one or more combining characters into the corresponding precomposed characters.
Aside: in general terms, this of course goes against the grain of the discussion above, but it is the general recommendation.
Assigning to Stefan for review.
fixed it in accordance to the request