what are the main differences between spec 2 and alpha?

mainly how they handle aggregates of attributes. You only specify required/optional attributes at the context of the usage site

‘bags’ of attributes are now the primary way of mapping domain entities in spec2, they don’t talk about requirements

also some programmatic convenience improvements

of course I am oversimplifying

https://github.com/clojure/spec-alpha2/wiki/Differences-from-spec.alpha

and https://github.com/clojure/spec-alpha2/wiki/Schema-and-select

hello, with spec1, I'm having issues to define a spec that is not loaded, I get Unable to resolve spec: ..., consider this is file B, the spec in question is defined by A, but A requires B, so I can't load A before loading B (circular reference), with that constraint, is there a way to tell spec to ignore that the spec is not defined at that time?

most spec types delay resolving references so don't run into this, but there are a few that don't

for those that don't, you'll need to break that cycle somehow

thanks, I got rid of the cycle by moving those spec definitions to a new namespace (since I we can define things from other namespaces, so still have the same name, just defined in a different file)

What's the best way to restrict a string to a min/max length, so that the generator still works for it ?

I don't think there is a general way to do that without writing a generator

(and to do that, I would generate strings of length up to (max-min) and a string of length min, then concat)

Hum, alright, I'll use a custom generator then

thx