'Umlaut correction' does not provide the right feedback
While I'm using the German umlaut as an example here, this also applies to the accents in Spanish. I assume it's the same in other languages.
If you get a question wrong and use the German shorthand for an umlaut (e.g. ue instead of ü), the parser picks this up as an error. Occasionally the actual error is altogether ignored in the feedback.
With the latest update, there is sometimes no feedback at all with regards to umlaut positioning: