I recognize this issue as something we found out already before. After a little bit of digging, it turns out that this issue was already uncovered last May:
It doesn’t seem like the issue has been fixed yet. I will check with iFastNet if there is anything new to report regarding this issue.
But it’s possible that this issue is low on the backlog because very few websites are actually affected by this. So you may need to keep patching the CMS to work around this issue for now.