Encoding not handled correctly for importing an existing bpmn file

I have tried to import an existing bpmn file using the modeler over bpmn.io - open.
If the file has UTF-8 encoding. Everything works just fine.

In case of windows-1252 encoding. The text is not displayed in the right way.

<?xml version="1.0" encoding="windows-1252"?>
<bpmn:definitions ..>
    <bpmn:task id="Task_08i9uk4" name="Überblick verschaffen">
      <bpmn:documentation>Überblick über alle aufgaben verschaffen</bpmn:documentation>
      <bpmn:extensionElements />


Überblick verschaffen.bpmn (3.1 KB)

This is correct. We treat all files as UTF-8. As such, we ignore this processing instruction, too:

<?xml version="1.0" encoding="windows-1252"?>

Which editor did you export from? Why does it use windows-1252 instead of UTF-8, the de-facto encoding standard?