Basically, Reflex should implement it's own version of Node.normalize(). This should only be applied in the initial primitive -> meta translation phase. We have to be careful not to over-optimize here because some of these ContentMeta objects could have been added to the tree as a result of a recurrent function. (In this case, it needs to remain a separate object so that it can be pulled out later)
Basically, Reflex should implement it's own version of Node.normalize(). This should only be applied in the initial primitive -> meta translation phase. We have to be careful not to over-optimize here because some of these ContentMeta objects could have been added to the tree as a result of a recurrent function. (In this case, it needs to remain a separate object so that it can be pulled out later)