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
Failure logs make it possible to debug failures of visual tests when they happen remotely on Travis or on CRAN checking machines. The failures are logged to a .Rout.fail file that gets displayed by Travis or on the CRAN page.
Right now vdiffr is set up so that by default it only pushes the failure log file when the NOT_CRAN environment variable is not defined. This was meant to avoid creating a log file when testthat is run interactively with devtools::test() and only create it when testthat is run via R CMD check.
However this prevents the log file to be created when NOT_CRAN is manually set on Travis. It would make sense to create the failure log when the CI environment variable is set by Travis, Appveyor, etc.
Failure logs make it possible to debug failures of visual tests when they happen remotely on Travis or on CRAN checking machines. The failures are logged to a .Rout.fail file that gets displayed by Travis or on the CRAN page.
Right now vdiffr is set up so that by default it only pushes the failure log file when the
NOT_CRAN
environment variable is not defined. This was meant to avoid creating a log file when testthat is run interactively withdevtools::test()
and only create it when testthat is run via R CMD check.However this prevents the log file to be created when
NOT_CRAN
is manually set on Travis. It would make sense to create the failure log when theCI
environment variable is set by Travis, Appveyor, etc.The function to modify is in utils.R:
vdiffr/R/utils.R
Line 92 in ef37e2b
is_checking()
predicate should be replaced by something that also checks forCI
.The text was updated successfully, but these errors were encountered: