Description Benjamin Schwertfeger 2010-08-05 04:16:44 EDT

The XtextResource load's the document input stream without encoding in the reparse method and call doLoad afterwards. This leads to broken encodings in the model. This is only visible, if you look at the document in memory, because the editor won't show the changes. The correct way would be an InputStream with encoding inside the reload method.

In my use case is a special action, which uses the current model in memory, transforms it and write to a different file. You can observe the wrong encoding, if you set a breakpoint in your model implementation files, where names are set, and clean your workspace to trigger reparse.