so let's say I am enriching a data structure with chained transformations (`map`-ing or specter-ing)...and I want to spec every substep. For instance the first step has {:family {:name "bla"}} the second step will have {:family {:name "bla" :id :sda}}.

how can I spec the second step after having already spec-ed the first as:

(s/def :family/entity (s/keys :req-un [:family/name]))
(s/def :family/family :family/entity)
(s/def :root/entity (s/keys :req-un  [:family/family]))

ups

now it should be ok

so basically I'd need to add another spec adding :family/id but I can't see how to do it without overriding the :family/family spec

ok I could change the namespace probably

Can you s/merge the :family/family spec with a new one which only specifies the new element?

maybe 😄 let me try that

but then I still need to change the namespace in the register

You could eval the s/def with a modified name…although having been doing a lot of it, I’ll warn you that that’s a route that can get reeeaaally sticky pretty fast. & there seem to be issues with evaling s/merge (that’s where I stopped today, had to call a halt. In fact, crap, what am I doing here? Going AFK immediately 😉 )

it works, but now for each transformation I need:

(s/def :entity-with-id/family (s/merge :family/entity (s/keys :req-un [:family/id])))
(s/def :entity-with-id/definitions (s/keys :req-un [:entity-with-id/family]))

so the namespace becomes the name of the transformation I want to target

hey all. is there a handy spec i can use for things that can be passed to s/valid? for testing?

i.e. name-spaced keywords, single-arity fn’s and specs I guess

aren't specs just either of the first two?

is there something like “merge, but make all the keys optional”?

or i guess optionally-merge?

not sure what i mean / want 😛

also thinking aloud: i kinda want a predicate that is basically any? but means “not yet spec’d"

@bbloom (def not-specd-yet? any?)