Skip to content

Latest commit

 

History

History
55 lines (40 loc) · 2.6 KB

etherpad.md

File metadata and controls

55 lines (40 loc) · 2.6 KB

Etherpad notes, archived from https://etherpad.mozilla.org/5c1KfAz3ME

Distribution logistics

Jobs:

  • Common taxonomy (master master, us/them, different repos etc)
  • Set up proof of concept for travis lesson checking
  • Documentation for rrhack organisers, bootcamp organisers, students

Travis

Travis for R. This is the repo that has code https://github.com/craigcitro/r-travis This is the recipes, which is super useful for designing new configurations. https://github.com/craigcitro/r-travis/wiki/Recipes

Here is the configuration file for a knitr document https://github.com/csgillespie/travis-examples/blob/knitr/.travis.yml

The one missing bit is that we really don't want a DESCRIPTION file in the repo, because it's very prominently named and will confuse the students. But I think we can solve this with a .description file that travis will rename.

Builds will be listed here: https://travis-ci.org/Reproducible-Science-Curriculum/rr-logistics/ (travis: unix build) https://ci.appveyor.com/project/richfitz/rr-logistics (appveyor: windows build) (travis has a long queue at the moment)

Code for buttons: Build Status Build status

This is quite likely to change in a little bit, but it will only get simpler and the current solution should work for a while https://twitter.com/GaborCsardi/status/542712472155983873

Ideas for things to test

  • Does a knitr document compile?
  • Does a derived data set match the master data set? (This is relevant for org1 vs org2)
  • Is a directory of figures set up?
  • Pull request example: Reproducible-Science-Curriculum/rr-logistics#2 - left off a closing backtick

Unresolved

  • It would be nice if Rscript could run in verbose mode (done)
  • Need to run travis over every subdirectory on the combined repo. (done)
  • Will need to manually enable travis on the new combined repo (not sure this is doable via api, but is not that big a deal)
  • Run each script with try, and throw at the end if any failed. This will pick up more errors.
  • Skip a directory if .ci_tests.R is not present
  • Make sure that the system calls actually fail - intern=TRUE might be enough.
  • Get links to the website running from each repo
  • Build a table of contents for the metarepo home page

Instructor materials - gh-pages

index.md file http://reproducible-science-cirriculum.github.io/rr-logistics