Just guessing from past experience in the Clojure internals:
Instrumenting protocol methods it likely possible, but unlikely to be worth
the effort.
There are already a bunch of places where protocol functions are different.
The accepted convention is to always wrap a protocol method with a nor
On Sunday, June 5, 2016 at 8:23:10 PM UTC+2, red...@gmail.com wrote:
>
> http://dev.clojure.org/jira/browse/CLJ-1941 has some discussion about
> places where instrumenting won't work.
>
>
Does this mean that this is a bug or something crucial related to protocol
design is preventing this ever be
http://dev.clojure.org/jira/browse/CLJ-1941 has some discussion about
places where instrumenting won't work.
On 06/05/2016 09:57 AM, Claudius Nicolae wrote:
> It seems that protocol fns don't participate in s/fdef specfications. It
> would be nice they were. Sample:
>
> (ns sample
> (:r
It seems that protocol fns don't participate in s/fdef specfications. It
would be nice they were. Sample:
(ns sample
(:require [clojure.spec :as s]))
(s/instrument-all)
;; Spec a fn
(s/fdef f
:args (s/cat :i integer?)
:ret integer?)
(defn
