Moving to sf

[boshek]

Sorry this isn't the cleanest pull request. I have had to fiddle with travis quite a bit to get this to work.

Have revised the approach to close #45 by replacing sp with sf (and geosphere) and making the installation of those packages optional. sf is becoming the standard spatial package so supporting it moving forward i think makes more sense that relying on sp.

I've added that to suggests. I have also revised the travis.yml file to remove osx. As far as I can tell most prominent R packages don't do CI on Apple:

• https://github.com/tidyverse/ggplot2/blob/master/.travis.yml
• https://github.com/tidyverse/dplyr/blob/master/.travis.yml
• https://github.com/RcppCore/Rcpp/blob/master/.travis.yml

Let me know if you'd prefer to make the commit tree cleaner.

[codecov-io]

Codecov Report

Merging #47 into steffi_exp will increase coverage by 0.04%. The diff coverage is 100%.

@@              Coverage Diff              @@
##           steffi_exp     #47      +/-   ##
=============================================
+ Coverage       93.86%   93.9%   +0.04%     
=============================================
  Files               5       5              
  Lines             505     509       +4     
=============================================
+ Hits              474     478       +4     
  Misses             31      31

Impacted Files	Coverage Δ
R/stations.R	99.09% <100%> (+0.03%)	: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 ccbabcc...54b8e33. Read the comment docs.

[boshek]

Based on this discussion I don't think the failures here are consequential: https://community.rstudio.com/t/r-devel-error-on-travis-ci-and-appveyor/2708/4

[boshek]

Yep you are right here. I'll just close the PR.