Upon a screenshare with Fox, we discovered there were \n and null characters in the incoming bytes that were not being matched and consumed by the serial data source, which caused the data source to seize up when these characters started clogging the buffer. A combination of rewriting the regex and preventing the nulls from transmitting brought this issue resolution!
Edit: A similar issue happened again after I proclaimed resolution. This led me to a concurrency bug in the virtual serial server socket.
Perhaps, I can't say what the original intention was. To me, it seems like it was for a new meta point, where you enabled it at some time and it begun to acquire data, then you can generate history without having to figure out when you enabled the point.
Yes, I can confirm the problem. I found the cause and fixed it, which will be in the next 3.5 release. Thanks for bringing this to our attention! The problem is that the active / inactive scripts are not being set on the import, which means you cannot create a new event handler that is using scripts or modify an existing script through the configuration import / export.