Conversation
|
This is how benchmark results would change (along with a 95% confidence interval in relative change) if e608c15 is merged into master:
|
|
@paul-buerkner Dependency resolution fails on Ubuntu 3.6 since a minimum version of randomForest is pinned which requires R 4.1 (see logs). Can the versioning constraint on randomForest be relaxed? Could vignettes be skipped on 3.6? |
|
This is how benchmark results would change (along with a 95% confidence interval in relative change) if b71f9e1 is merged into master:
|
|
This is how benchmark results would change (along with a 95% confidence interval in relative change) if eef4937 is merged into master:
|
|
You can just merge these github action PRs without my checks if you have the github permissions for that |
|
This is how benchmark results would change (along with a 95% confidence interval in relative change) if b8b7070 is merged into master:
|
This reverts commit 73ae92c.
|
This is how benchmark results would change (along with a 95% confidence interval in relative change) if 7889738 is merged into master:
|
Summary
Fix failing R CMD Checks by swapping to r-lib GHA steps to handle sys/R dependency resolution. Also moves from bespoke steps to r-lib or native GHA steps/config.
For Ubuntu with R 3.6, since the suggested package randomForest requires 4.1, this PR selectively deletes lines in the action to not need randomForest (in example and DESCRIPTION). Pipes in tests are also removed to run on R 3.6
Copyright and Licensing
By submitting this pull request, the copyright holder is agreeing to
license the submitted work under the following licenses: