Test cases

[stephlocke]

Here's a list of Rmd files from over the years that'd be good for testing!

• [x] https://github.com/reconhub/learn/tree/master/content/post
• [x] https://github.com/lockedata/trainingportal/tree/master/handouts
• [x] https://github.com/lockedata/trainingportal/tree/master/handouts/book1
• [x] https://github.com/lockedata/trainingportal/tree/master/handouts/book2
• [x] https://github.com/lockedata/brentsbiz_study

Hopefully some folks will be to suggest others

[maelle]

https://github.com/reconhub/learn/pull/31

[maelle]

Notes of things I discovered whilst testing:

• I was parsing the headers too much, now I rely on knitr code and it all goes much better.

• There can be spaces in chunk labels so when I split the label by space to find the language and label, I lost part of the label. Now the code replaces the first space with a slash and I split by slash.

• I hadn't taken into account the case when the Rmd doesn't contain any chunk, now I do.

• If a Rmd file already contains duplicate names, I get an informative error message, thanks previous-me.