A sketch of my tree looks like this:
I'm trying to read this as:
Obviously, I had to write a bug report. I'm guessing that the second and third error messages are just error handling gone wrong, and the real issue is the first error. Using TTree::Print and the tree-viewer I can verify that the values are written correctly.
Now, it shouldn't be too hard to translate a C-Style array into whatever type you want to use for the TTreeReaderArray, so I would consider any inability to do so a serious shortcoming [*], but that's not what the error message is telling me.
What's really confusing is the last part of the first message, where I'm asked to "access one of its siblings". Is that Japanese-type politeness (i.e. instead of saying "don't do this" it says "please do something entirely unrelated") or is that pointing me to a different way of achieving my ends?
[*] I'm aware of the perceptive bias that encourages me to assess this shortcoming as serious