Patches to add code coverage support in the R interpreter.
This repository contains patches of the R interpreter source code to add code coverage support. It allows to know which exact lines of source code were exercized/hit/covered while running some piece of code (usually a test suite). The easiest way to test it is using our docker container: https://github.com/quartzbio/r-coverage-docker
As stated above, you can just use the docker container. To patch it manually:
- choose the appropriate patch corresponding to the version of the R interpreter you are going to use.
- get the source code of the R version (download and untar)
- patch the source code (patch)
- configure the source directory (./configure)
- compile and install (make; make install)
The project is organized in sudirectories per supported R version, e.g. r302 for R version 3.0.2. The above steps are somewhat automated using a Makefile (you need probably GNU make). For example, to download, untar, patch, configure, make and install R-3.0.2:
git clone https://github.com/quartzbio/r-coverage-patch.git cd r-coverage-patch/r302/ # download, untar and patch make apply_production_patch # install in ./local/ make install # Or install in a custom place make install PREFIX=/whatever/ # run it ./local/bin/R
I added a condition in the internal C function getSrcref(), that records the line numbers if the code coverage is started (via Rcov_start()). The overhead should be minimal since for a given file, subsequent covered lines will be stored in constant time. I use a hashed env to store the occurrences by file.
I added two entry points in the utils package (Rcov_start() and Rcov_stop())
viewing the main patched file: eval.c
For convenience (Caution: it can be outdated), I included the main patched file in the repo, you can view it here: https://github.com/quartzbio/r-coverage-patch/blob/master/r302/patched_files/eval.c
Start the patched R interpreter, then:
library(devtools) pkg <- download.packages('testthat', '.', repos = "http://stat.ethz.ch/CRAN") untar(pkg[1, 2]) Rcov_start() test('testthat') res <- as.list(Rcov_stop()) print(res)