[R-pkg-devel] package daewr does not pass incoming checks automatically

Duncan Murdoch murdoch@dunc@n @end|ng |rom gm@||@com
Tue Mar 10 18:02:02 CET 2020


On 10/03/2020 9:48 a.m., John Lawson wrote:
> The package daewr contains the data sets and functions shown in the book
> "Design and Analysis of Experiments with R" published by CRC press. It has
> been on CRAN since the book was first published in 2015.  Recently I
> included a new function, but have not been able to include the modified
> package on CRAN.
> 
> I get the message from Debian:
> 
> * DONE
> Status: 1 WARNING, 1 NOTE
> 
> Current CRAN status: WARN: 1, OK: 12
> See: <https://CRAN.R-project.org/web/checks/check_results_daewr.html>
> 
> Version: 1.1-8
> Check: package dependencies, Result: WARNING
>    Requires orphaned package: ‘BsMD’
> See: <https://www.r-project.org/nosvn/R.check/r-devel-linux-x86_64-fedora-clang/daewr-00check.html>
> 
> 
> I do not get this warning from the Windows:
> 
> Although the package BsMD is orphaned, it is still available on CRAN and
> all examples in the book that use functions in daewr that call functions
> from BsMD work as they always have. The new function I have added does not
> require BsMD. Is there any way I can upload my modified package to CRAN so
> that the new function is available?  Any help would be greatly appreciated.
> 

I suspect the logic is this:

Even though BsMD checks OK now, eventually some change in R or the check 
process will likely require changes to it.  However, since it is 
orphaned, there is nobody to ask to make those changes, so other 
packages should be discouraged from depending on it until a new 
maintainer steps forward.

So you have three choices:

1.  Volunteer to maintain BsMD, and take on the responsibility of 
modifying it in the future.

2.  Stop depending on it now, so that when it eventually breaks, your 
package won't be affected.

Either one of those choices requires some extra work from you, but it's 
not an emergency, as it would be if BsMD was suddenly unavailable.

Your third choice is less work, but less satisfying:

3.  Don't include the new function until someone else steps forward to 
maintain BsMD, or you are forced to make changes for some other reason.

Duncan Murdoch



More information about the R-package-devel mailing list