predict_gnls with interval="prediction" hangs asking for grouped data

[piodag]

Hi,
I'm trying to use predict_gnls() on a gnls model that contains two factors with interactions and a varIdent variance. With the option interval="confidence" I get the CI with a simple newdata data.frame. If I use the same data.frame with the option "prediction" the function stops asking for grouped data. But even faking some data and using an object created with groupedData() do not fix the problem. Am I missing some very basic concept? Or is it a bug? Sorry to ask it here. I have no idea how to be able to reach out for a reply.
TIA
Giorgio

[femiguez]

Hi @piodag This is the perfect place to ask. I think I understand the issue and will look into it. What is usually really helpful is to post here a reproducible example, so it is more clear. I will get back to you after I investigate, but an example would be appreciated. Best, Fernando

[femiguez]

Please submit an example. It might be a bug in the nlme package