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
Hello, thank you to the maintainers and contributors that creating a wonderful package. I was hoping we could possible start a discussion regarding best practices in using this package, specifically expect_doppelganger() and it consistently failing on a particular OS environment. In my use case, I am using github actions to automatically run R CMD check for various OS environments:
Lately, only the ubuntu-latest, r:'devel' environment is failing consistently, and after downloading the output and reviewing the differences with testthat::snapshot_review() I see that the differences are very minuscule, and often related to how paths were drawn, not obvious to the human eye.
My concern here is largely due CRAN's automated checks. Would this be one of those things in which it would be recommended to skip those specific tests with testthat:skip_on_cran(), at least until the CI environments start to pass again?
Or is there a way to have the vdiffr::expect_doppelganger() make new snapshot svg files for specific OS environments and r version combinations?
Any advice is greatly appreciated
The text was updated successfully, but these errors were encountered:
Note that the graphical snapshot tests from {vdiffr} are skipped by default, and so you don't need to add skips for {testhat} unless you've overridden this argument:
OS-specific failures
This is definitely a trickier issue to deal with1:
Snapshot testing for text/file in {testthat} offers a variant argument to deal with this issue.
Snapshot testing for Shiny apps in {shinytest2} also offers a function for the same.
AFAICT, {vdiffr} currently doesn't allow a straightforward API to create variants of snapshots, and it would indeed be a nice addition.
Footnotes
I've more detailed thoughts on snapshot testing in these slides, in case you are interested. ↩
Hello, thank you to the maintainers and contributors that creating a wonderful package. I was hoping we could possible start a discussion regarding best practices in using this package, specifically
expect_doppelganger()
and it consistently failing on a particular OS environment. In my use case, I am using github actions to automatically run R CMD check for various OS environments:Lately, only the ubuntu-latest, r:'devel' environment is failing consistently, and after downloading the output and reviewing the differences with
testthat::snapshot_review()
I see that the differences are very minuscule, and often related to how paths were drawn, not obvious to the human eye.My concern here is largely due CRAN's automated checks. Would this be one of those things in which it would be recommended to skip those specific tests with
testthat:skip_on_cran()
, at least until the CI environments start to pass again?Or is there a way to have the
vdiffr::expect_doppelganger()
make new snapshot svg files for specific OS environments and r version combinations?Any advice is greatly appreciated
The text was updated successfully, but these errors were encountered: