We analyzed this issue and found that it was caused by two bugs actually:
First, the TestRail importer produced invalid documents in very rare cases with lists.
Second, when an invalid document is first edited, it’s normalized and by default this process thrww away parts that are not allowed. We now try to retain as much information as possible while still normalizing the document.
We tried to fix both issues with our latest release, please give it a try. Thanks!
See the release announcement here.
Best regards,
Alex