Clojurians Log v2
pathom
I've been wondering bit about the ergonomics of Pathom, particularly the way it breaks up its namespaces. I find that when I start to work with it I end up needing to require several pathom namespaces and end up regularly using just 1-2 functions/keywords per namespace; By now I have these namespaces somewhat memorized and think I understand the conceptual seperations, but still end up just copy/pasting a big chunk of requires when I start exploring the usage of pathom in a project. I've just started experimenting with making my own namespace that just wraps a large amount of the pathom functionality so that it can all be used with a single require. This was surprisingly easy to do which suggests this whole thing isn't much of a problem. Though I'm not sure yet if I want to build up my codebase around these wrappers instead of just pathom proper. I'm curious if anyone else has thoughts on this. I know that in a way this is all minutia. I'm really grateful for pathom btw!
at work we recently had the same idea to make a wrapper namespace for both pathom and fulcro--so +1 for that approach.
hello, the separation of namespaces is done for semantic reasons first at Pathom 3, because a lot of times the place where you create resolvers/mutations you may need operations, while the configuration of runtime usually lives in another namespace. so its an organizational-first principle to me there.
I guess the friction/pain is a bit tied to the user editor, I myself dont have problems with multiple namespaces because I just write things like pci/regi... and let my editor (Cursive) auto-import the thing for me. I think its similar to how people do JS/TS nowadays (I had done a bit, and you just type the things and let editor auto-import it).
that said, I think its totally valid to make your own uber namespace with all the things, I guess this is more useful for demo/playground than in full-size projects, given IME those tend to sparce out the usage of the namespaces
Thanks, makes sense