wontrun

The goal of {wontrun} is to test the longevity of old code for the R programming language by making it easy to run examples from archived package sources on current versions of packages.

Installation

{wontrun} is not available on CRAN, but you can install the development version like so:

# install.packages("devtools")
devtools::install_github("b-rodrigues/wontrun", ref = "master")

Example

Get the archived sources for the {dplyr} package:

library(wontrun)
dplyr_sources <- get_archived_sources("dplyr")

This is what this looks like:

dplyr_sources
#> # A tibble: 39 × 5
#>    name  version       url                             last_modified       size 
#>    <chr> <chr>         <chr>                           <dttm>              <chr>
#>  1 dplyr dplyr_0.1.1   https://cran.r-project.org/src… 2014-01-29 21:24:00 530K 
#>  2 dplyr dplyr_0.1.2   https://cran.r-project.org/src… 2014-02-24 16:36:00 533K 
#>  3 dplyr dplyr_0.1.3   https://cran.r-project.org/src… 2014-03-15 00:36:00 535K 
#>  4 dplyr dplyr_0.1     https://cran.r-project.org/src… 2014-01-16 22:53:00 2.7M 
#>  5 dplyr dplyr_0.2     https://cran.r-project.org/src… 2014-05-21 08:20:00 577K 
#>  6 dplyr dplyr_0.3.0.1 https://cran.r-project.org/src… 2014-10-08 05:21:00 629K 
#>  7 dplyr dplyr_0.3.0.2 https://cran.r-project.org/src… 2014-10-11 07:43:00 628K 
#>  8 dplyr dplyr_0.3     https://cran.r-project.org/src… 2014-10-04 06:39:00 629K 
#>  9 dplyr dplyr_0.4.0   https://cran.r-project.org/src… 2015-01-08 11:12:00 870K 
#> 10 dplyr dplyr_0.4.1   https://cran.r-project.org/src… 2015-01-14 07:15:00 870K 
#> # … with 29 more rows

Let’s also choose how many cores we want to use to run the examples, as this is a time-consuming operation:

cl <- 6

We can now run the examples contained in the archived packages with the wontrun() function:

# By default, wontrun() will only consider the earliest package in a given year
dplyr_runs <- wontrun(dplyr_sources, ncpus = cl)

The package provides a decode_wontrun() function to summarise the results:

decode_wontrun(dplyr_runs)
#> # A tibble: 5,328 × 7
#>    name  version   last_modified       scripts_pa…¹ runs         classes message
#>    <chr> <chr>     <dttm>              <chr>        <list>       <chr>   <chr>  
#>  1 dplyr dplyr_0.1 2014-01-16 22:53:00 /tmp/Rtmp8X… <named list> list    ""     
#>  2 dplyr dplyr_0.1 2014-01-16 22:53:00 /tmp/Rtmp8X… <pckgStrM>   packag… "Loadi…
#>  3 dplyr dplyr_0.1 2014-01-16 22:53:00 /tmp/Rtmp8X… <smplErrr>   simple… "could…
#>  4 dplyr dplyr_0.1 2014-01-16 22:53:00 /tmp/Rtmp8X… <pckgStrM>   packag… "Loadi…
#>  5 dplyr dplyr_0.1 2014-01-16 22:53:00 /tmp/Rtmp8X… <pckgStrM>   packag… "Loadi…
#>  6 dplyr dplyr_0.1 2014-01-16 22:53:00 /tmp/Rtmp8X… <lfcycl_s>   lifecy… "src_s…
#>  7 dplyr dplyr_0.1 2014-01-16 22:53:00 /tmp/Rtmp8X… <named list> list    ""     
#>  8 dplyr dplyr_0.1 2014-01-16 22:53:00 /tmp/Rtmp8X… <pckgStrM>   packag… "Loadi…
#>  9 dplyr dplyr_0.1 2014-01-16 22:53:00 /tmp/Rtmp8X… <smplErrr>   simple… "could…
#> 10 dplyr dplyr_0.1 2014-01-16 22:53:00 /tmp/Rtmp8X… <pckgStrM>   packag… "Loadi…
#> # … with 5,318 more rows, and abbreviated variable name ¹​scripts_paths

Thanks

• Thanks to Dirk Eddelbuettel for the idea!
• Thanks to Deemah for suggesting the name of the package!
• Thanks to Jenny Bryan for suggesting using {callr} which I use in this package.
• Thanks to David Hood for suggesting the term code longevity.
• and thanks to everyone on Mastodon for the helpful discussions!