Clean up and add docstrings

[palday]

No description provided.

[codecov]

Codecov Report

Attention: Patch coverage is 83.33333% with 3 lines in your changes missing coverage. Please review.

Project coverage is 80.69%. Comparing base (b6f4f19) to head (2bd1599).

Files	Patch %	Lines
src/types.jl	70.00%	3 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #539      +/-   ##
==========================================
+ Coverage   80.65%   80.69%   +0.03%     
==========================================
  Files          26       26              
  Lines        1701     1704       +3     
==========================================
+ Hits         1372     1375       +3     
  Misses        329      329              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[dmbates]

Thanks for cleaning this code up. I have a couple of minor suggestions.

I noticed that the tests on nightly are failing, apparently because of a change in the REPL implementation. I assume the failure is unrelated to these changes. Is that the case?

[dmbates]

I'll let @palday handle the merging of the changes in base into this branch. Those kinds of git operations make me want to run away and hide.

[ararslan]

I assume the failure is unrelated to these changes. Is that the case?

Yes, nightly has been failing on all recent PRs.

[ararslan]

This isn't functionally the same as what was written but I'm not sure whether what was written before was intentional. It was implicitly relying on convert(Ptr{S}, p::Ptr{SymSxp}) but I don't know enough about how this package works to know whether that's a sensible conversion to do. With this change, RObject{S}(::Ptr{SymSxp}) for S != SymSxp will be a MethodError, won't it?

[palday]

Yeah, I think it will be a method error, but I think that having a method error there makes sense. The usual call would be RObject(p), which invokes this definition:

RObject(p::Ptr{S}) where S<:Sxp = RObject{S}(p)

and thus would always use the narrowest possible construction.