revdepcheck
revdepcheck copied to clipboard
r-lib
R package reverse dependency checking
When considering an API breaking change, it is useful to be able to empirically determine the impact by comparing revdeps from a branch with revdeps on master.
It would be nice to see the time taken for each package as the check is going. Something like ``` ✔ RTCGA 1.6.0 ── E: 2 | W: 0 |...
Older code that sets up various test scenarios, might be useful: https://gist.github.com/krlmlr/8ba1f9a62bf75eb8afb2441b997e8319.
E.g. if the output is `/Users/jhester/Dropbox/projects/devtools/revdep/checks/annotatr/new/annotatr.Rcheck/00_pkg_src/annotatr/R/visualize.R` it should be converted to something like `.../revdep/checks/annotatr/new/annotatr.Rcheck/00_pkg_src/annotatr/R/visualize.R`. The rest of the path is not really relevant to the error report. Should remove the...
Could be a global option. This would help when the package root is within Dropbox or equivalent. The large amounts of created files causes intense CPU usage by Dropbox.
For troubleshooting purposes, having `revdep_check()` report: ```r if (file_test("-f", p
I'd like to be able to send a blanket email to all maintainers. This is different than to just the failures. Sometimes a package doesn't break the standard R tests.
From https://github.com/r-lib/rcmdcheck/issues/71: > It seems, `revdep_check` ignores argument `libpath` when set to a character, e.g. `"C:/Users/username/rlibs"`. What happens is that it tries to write to the default library. So, I...
Otherwise packages such as _xgboost_ will happily use all CPUs available. https://www.ibm.com/support/knowledgecenter/SSGH2K_12.1.0/com.ibm.xlc121.aix.doc/compiler_ref/omp_thread_limit.html