Beyond inline expectations

When you want to test

• large/complicated output,

• binary files like images,

• error messages…

Think about snapshot testing!

From testthat third edition.

Another testthat trick

Custom skippers to run some tests only where you want.

Some built-in skippers: operating system, CRAN, etc.

Demo 🚀

What we just saw

• Snapshot testing

• Custom skippers

What about real requests?

Fixtures are great but what if the API output changed & we didn’t notice? 😱

Let’s schedule runs of continuous integration, where we turn off vcr.

Demo 🚀

How to get better at writing CI workflows?

Get inspired by others!

LIFE HACK: My go-to strategy for getting Travis builds to work is snooping on *other* people's .travis.yml files. Shoutout today to the tidyr .travis.yml for solving my problem! #rstats 🙌

— Julia Silge (@juliasilge) December 12, 2019

How to get better at testing?

• Also get inspired by others!

• Read testthat docs, the HTTP testing in R book, etc.

HTTP testing

• vcr handy for usual workflows, and highly configurable.

• Other HTTP testing packages: httptest, webfakes.

• For databases, see dittodb!

Thank you, useR! team & participants!

Thanks in particular to Heather Turner (technical helper), Liz Hare (useR! accessibility team) and Yanina Bellini Saibene (global coordinator). And thanks to all other useR! volunteers!

Questions

• Right now if we have time.

• In the lounge during the next days but asynchronously.

• rOpenSci forum, RStudio community forum.