satisfies? finds protocols that have no methods

[lilactown]

Please answer the following questions and leave the below in as part of your PR.

• I have read the developer documentation.
• This PR corresponds to an issue with a clear problem statement.
• This PR contains a test to prevent against future regressions

---

Fixes #785

The crux of the issue is that the way that satisfies? currently turns a type into a symbol,

sci/src/sci/impl/protocols.cljc

Lines 245 to 251 in 3e068bb

	(let [t (cond
	#?(:clj (class? t))
	#?(:clj (symbol (.getName ^Class t)))
	(instance? sci.lang.Type t)
	(symbol (str t))
	:else t)]
	(contains? sats t)))))

does not match the way that extend-type and extend-protocol do when adding the metadata to the protocol

sci/src/sci/impl/protocols.cljc

Lines 232 to 235 in 3e068bb

	(clojure.core/alter-var-root
	(var ~proto) update :satisfies (fnil conj #{})
	(symbol (str ~atype)))
	~@(process-methods ctx atype meths pns extend-via-metadata)))) proto+meths))))

This PR attempts to unify the way we coerce the type to a symbol between extend-type, extend-protocol and satisfies?.

In addition, it also exposes a new var clojure.core/type->str. The rationale for it is we need to use the same machinery in two different contexts: a macro context in extend-type and extend-protocol, and at runtime in satisfies?. I could emit the code directly in the macro context instead, but I thought it would be better for maintenance purposes to have everything refer to a single function.

There also already exists a type->str function in CLJS, so the only maybe-surprising thing for consumers is that there now exists a clojure.core/type->symbol function in SCI that doesn't exist in JVM Clojure.

[borkdude]

so the only maybe-surprising thing for consumers is that there now exists a clojure.core/type->symbol function in SCI that doesn't exist in JVM Clojure.

Perhaps move this to a "private" namespace. There are some of those in namespaces, perhaps there's already a protocols-specific one.

[lilactown]

@borkdude I've update this PR to use just type->str without converting to a symbol when extending a protocol and checking for satisfies?, as well as exposed the type->str function via a private sci.impl.protocols SCI ns rather than from clojure.core

[borkdude]

@lilactown Excellent!