When digitizing a word document into a Xerte quiz i added a question which had the square root of a number in it. When copying this over, Xerte did not recognise the character and showed the white square unicode character (□). After publishing these changes and closing the project the entire thing became corrupt. I could still open it, but the pages were completely blank and nothing allowed me to get all the originally added content back. Previewing the project also resulted in a blank screen.
Removing the white square unicode characters resolves the issue and obviously you don't want these in the final project anyway. But work can easily be lost if you're copying content and not paying attention or just closing the project and working on it later.