Use readtext for other extensions

[adam3smith]

The idea is to use readtext whenever it explicitly supports a format (list taken from their manual)

[greebie]

I think this accomplishes the task described in the heading, so I think we should be fine merging. However, I did test this with the following as a test file:

 <?xml version="1.0" encoding="UTF-8"?>
 <links>
 <link>https://www.google.com</link>
 <link>https://www.example.com</link>
 <notalink>NOT A LINK</notalink>
 </links>

and it returns an empty list, which means we end up with an

 Error in matrix(unlist(newlst), nrow = length(newlst), byrow = T) : 
  'data' must be of a vector type, was 'NULL'

Which should not happen anyway (if we have a 'NULL' it should quietly ignore the entry).

I suggest we merge this and add an issue either to include the empty link boilerplate (which really should go away anyway) or use a partial function (if R has that or a similar feature) to disinclude all null entries in the lapply statement. The problem with XML can be catalogued as part of #18 (part of the solution would be to try and improve the regex).

(If you agree just thumbs up and I'll merge).

[adam3smith]

might be worthwhile to just use read_xml and read_html for the respective files rather than doing crazy regex...

[greebie]

Weird - I got a broken merge warning in my email, but checks passed according to the website.

[greebie]

Problem fixed itself after reset. Seemed like a blip with the R package for Travis.