better role handling in coerce_authors_at_r()

[dpprdan]

This PR should improve the handling of roles by coerce_authors_at_r(). My main assumptions underlying the code:

• Author and Maintainer fields are not empty, as required by WRE (but not sure about Bioconductor for example)
• Author and Maintainer contain strings that can be parsed by as.person() (violated by packages with an Author file, e.g. igraph, so probably needs an appropriate error message)
• Only one Maintainer, as required by WRE (again dunno about Bioconductor)

closes #114

[codecov-commenter]

Codecov Report

Merging #115 (1f175a2) into main (df76490) will increase coverage by 0.10%. The diff coverage is 100.00%.

:exclamation: Current head 1f175a2 differs from pull request most recent head 0747526. Consider uploading reports for the commit 0747526 to get more accurate results

@@            Coverage Diff             @@
##             main     #115      +/-   ##
==========================================
+ Coverage   97.40%   97.50%   +0.10%     
==========================================
  Files          18       18              
  Lines        1154     1164      +10     
==========================================
+ Hits         1124     1135      +11     
+ Misses         30       29       -1     

Impacted Files	Coverage Δ
R/authors-at-r.R	91.98% <100.00%> (+0.79%)	:arrow_up:

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update df76490...0747526. Read the comment docs.

[gaborcsardi]

Thanks! I'll review and add this before the next release. Which might take a few months, because desc was just updated about a week ago. Hopefully that is OK.

[dpprdan]

No problem. I just added a few more changes, I hope the intent becomes clear from the code and the commit messages.

[gaborcsardi]

Thanks much and sorry for the long wait!