Need better understanding of janice_template_get_feature_vector semantics

[taa01776]

I can easily imagine that the feature_vector_type parameter would be a return value (that is, I give the caller a feature vector back, and tell him that it's full of floats), but that's not what's in the header file. It makes little sense for me for the caller to request a feature vector of a particular type: if my feature vectors are all full of signed floats, requesting uint8 doesn't make a whole lot of sense, and should just return an error--but there isn't really anything appropriate, except maybe "BAD_ARGUMENT". Or, I suppose, "NOT_IMPLEMENTED".
Please clarify.

[JordanCheney]

@taa01776 This is a bug. I just pushed 14631d5 to fix it. feature_vector_type should be an output parameter as you said.