I control Spotify and iTunes on my PC with a Harmony Elite remote and Girder, and I use Text To Speech (TTS) to announce the playlist name when I select one on the remote. Each playlist action contains a TTS node with the name of the playlist as the text.

So, I just upgraded my PC from Windows 7 to Windows 10 and one of the things that happened is that all my TTS nodes stopped working in Girder. A bit of investigations shows that the voice "Microsoft Anna" is no longer available, however replacing it in a TTS node with any available voice, e.g. "Microsoft Zira", works fine.

So I exported my Girder file and edited the .gml to replace "Anna" with "Zira". I then imported that file. None of my TTS nodes showed the change. Furthermore, any other change I made in the .gml file (e.g. a playlist name) does not show after importing the file. It looks like the import is not actually doing anything, even though there are no errors reported during the operation.

I'm using Girder 6.0.19.1

Am I missing something?