validate_tests.Rd
Will clone repo
at version
tag and run test suite. Package will be built
and installed (along with any missing dependencies) to a temporary library.
Then we run all tests for repo
, roll up successes and failures, and write
to a CSV file. Also writes the commit hash of the package and some other
system info to a JSON file.
Github organization that the repo is under
The name of the repo for the package you are validating
The version number of the package you are validating. This must correspond to a tag in the repo.
Domain where repo lives. Either "github.com" or "ghe.metrumrg.com", defaulting to "github.com"
File path to write out the test results to. Any extension
will be ignored and replaced with .csv
for the test output df and
.json
for the get_sys_info()
output.
The directory path to where the package has been cloned.
file.path(root_dir, repo)
should lead to the cloned repo that will be
tested.
Directory to write the test outputs to. Defaults to working directory.
Character vector of paths (relative to package root dir) to directories that contain additional tests to run
Whether to copy the TestName
column over to the
TestId
column when writing the output CSV. This exists to support legacy
issues and tests that don't use IDs. Starting with v1.0, mrgvalidate relies
on test IDs and, as a temporary compatibility kludge, will auto-generate
them if the TestId
and TestName
columns match. New tests and issues
should use test IDs.
Invisibly returns tibble of formatted test output. Note that this function is primarily called for the side effect of writing out the output to disk.
The package found at file.path(root_dir, repo)
will be built and installed
into a tempory library that is appended to the beginning of your
.libPaths()
, using devtools::install(..., build = TRUE, upgrade = "never")
. This will also install (into the temporary library) any
dependencies from the package that are not found in your .libPaths()
.