Ultimately I’m trying to add a branch to an :orn schema, and not seeing what I expected: Given

(def my-schema
  [:schema {:registry {"test" [:orn [:string-node :string]]}} "test"])

(mu/update-in my-schema [0 0] mu/assoc :int-node :int)
;; => [:schema {:registry {"test" [:orn [:string-node :string]]}} "test"]

(mu/get-in my-schema [0 0])
;; => [:orn [:string-node :string]]

(mu/assoc [:orn [:string-node :string]] :int-node :int)
;; => [:orn [:string-node :string] [:int-node :int]]

I’ve been doing a lot of this lately. maybe I can help

any kind of transform using registries is tricky. it’s much easier to deref the entire schema before starting any transforms

https://clojurians.slack.com/archives/CLDK6MFMK/p1696974522814249

once you have it fully local, then transforms are much easier to grok. does this help?

I ended up using m/form and a postwalk. I have control over the shape of the schema coming in, so it’s really just a superstition sort of thing.

it is for a convenience function for adding branches to my hiccup->html compiler https://github.com/escherize/huff#extendable-grammar 🙂

So, all good with this one?

:tuple is a fast fixed size vector. :cat is sequential schema with regex semantics. More powerful, bit slower.

the perf difference between the two:

(require '[criterium.core :as cc])

(let [valid? (m/validator [:* int?])]
  (cc/quick-bench (valid? (range 10)))) ; Execution time mean : 2.7µs

(let [valid? (m/validator [:sequential int?])]
  (cc/quick-bench (valid? (range 10)))) ; Execution time mean : 0.12µs

Thanks!