You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently the junit reporter writes a element for each expectation, with a classname attribute but no name attribute, leading to broken links in reporting interfaces like Jenkins:
The text was updated successfully, but these errors were encountered:
* Adds name attribute to <testcase> elements in junit output
* Updated junit test data
* Removed encoding attribute from xml tag in junit results
Perhaps a different version of libxml is being used than to
generate the original output.
* Fixed indenting
* Fix JUnit output
- "message" attribute mapped directly to expectation message
- failure/error body set to format(expectation)
This produces the same output as the summary reporter for
errors/failures.
* Fixed expected junit output paths
* Fixed (again) junit output
Matched the stack depth numbers to those output by the
summary reporter.
* Allow configuration of test_check() using options()
This adds two options:
* testthat.default_check_reporter: allows configuration of a different reporter
to use for test_check(). For example, 'junit'
* testthat.junit.output_file: specifies a file to which the junit xml should
be written.
These are needed so that Ci systems can run test scripts in R packages
without modification, but first set these options() so that test output will
be written to a known location where it can be read and parsed.
* Updated NEWS.md with changes to JUnitReporter and new options
Currently the junit reporter writes a element for each expectation, with a classname attribute but no name attribute, leading to broken links in reporting interfaces like Jenkins:
The text was updated successfully, but these errors were encountered: