Protocol fun documentation parity

[bobisageek]

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.

Aimed at directly addressing babashka/babashka#1340

• This PR contains a test to prevent against future regressions

Summary
Add an attribute map to the multimethods created for protocol functions. The attribute map is based directly on the signature map currently parsed by defprotocol. Also, add a test for doc output based on running the same protocol definition in clojure/jvm.

likely(?) discussion points

• arglists aren't enforced (as in arity checks) - this is also true for 'hand-crafted' arglists, but is a variance from protocols on cl/jvm, so I wanted to call that out in case these should be captured/rendered differently
• the "no arglist" test - I wanted to capture a variety of arglist/docstring combos for this test, and it occurred to me that it's technically possible (with SCI) to specify a protocol method with no arglists; the test is primarily there to ensure that these changes don't cause any undesired behavior around the metadata, but it might not be desirable to have a test scenario that's undesirable
• cross-linking issues in changelog - I only saw one other example of referring to an issue from the bb issue list, but it wasn't a link; not sure if there's a more preferred style for linking over (or maybe just eschew the linking altogether)

[borkdude]

I'll take a look at this tomorrow

[borkdude]

@bobisageek I didn't see anything that looked weird to me and tested it with bb master. Thanks for the improvements.