[R-pkg-devel] CRAN submission pretest warning about Rcpp version used to build dplyr

Anderson,Brooke Brooke.Anderson at colostate.edu
Thu Jul 13 18:58:29 CEST 2017


I got the following warning when submitting a package to CRAN that I am struggling to understand:


Warning: Installed Rcpp (0.12.12) different from Rcpp used to build dplyr (0.12.11).


Here is a link to the directory with the full logs from the CRAN pretest checks:

https://win-builder.r-project.org/incoming_pretest/170713_171433_noaastormevents_010/


I did not get this warning on any of the systems where I ran checks before submitting to CRAN

(I ran checks on my local Mac OS X install, which is R 3.4.0, on win-builder, and on Travis using the

“trusty” distribution, ubuntu 14.04.5 with R 3.4.0). Within the 00install.out file from the CRAN pretest

(available in the link to the full log posted above), there is the message:


Please reinstall dplyr to avoid random crashes or undefined behavior.


I am curious if reinstalling dplyr on my local system should fix this, given that I did not get this warning

or message on my local system, or if the warning might result from the versions of dplyr and Rcpp on

the system used by the CRAN pretest? The only other example I could find through googling

of someone having a similar problem is here: https://github.com/KopfLab/lans2r/blob/master/cran-comments.md.

It sounds like in that case the warning may have resulted from the configuration of win-builder at the

time and was not something that could be resolved by the package maintainer.


The package I submitted does not have a direct dependency on Rcpp but does import dplyr and other

tidyverse  packages. My local system (on which I built the package I submitted to CRAN) has

the latest CRAN versions of both Rcpp (0.12.11) and dplyr (0.7.1). The full source code for the

package I submitted is at  https://github.com/geanders/noaastormevents.


I have not had this warning or a similar in past packages I’ve submitted with dependencies on

dplyr. I looked through recent threads here on r-pkg-devel and could not find any related posts.

Does anyone have a suggestion for what’s causing this warning on the CRAN pretest and

how I can resolve it? Thanks very much in advance for any advice.


	[[alternative HTML version deleted]]



More information about the R-package-devel mailing list