[R-pkg-devel] how to address CRAN FAIL on r-devel-windows-x86_64-gcc10-UCRT?

David Winsemius dw|n@em|u@ @end|ng |rom comc@@t@net
Thu Apr 8 02:14:17 CEST 2021

On 4/2/21 1:08 PM, David Winsemius wrote:
> On 4/2/21 4:05 AM, Daniel Kelley wrote:
>> The "oce" package for oceanographic analysis is failing a check on 
>> the r-devel-windows-x86_64-gcc10-UCRT system (see
>> https://www.r-project.org/nosvn/R.check/r-devel-windows-x86_64-gcc10-UCRT/oce-00check.html). 
>> The problem seems to relate to the checking vignette outputs.
>> According to 
>> https://www.r-project.org/nosvn/winutf8/ucrt3/CRAN/checks/gcc10-UCRT/README.txt, 
>> that test setup is experimental.  Does that mean that I ought to I 
>> ignore the FAIL status for this machine, perhaps checking it once in 
>> a while to see if it changes?
> Hello Daniel;
> I'm not speaking for CRAN. I'm another maintainer whose current status 
> page has a FAIL for that platform. But the FAIL notation is due at 
> least in part to the fact that apparently the UCRT platform is less 
> frequently run on the full CRAN package holdings. The note refers to a 
> package that was updated a few days ago. The log file however has two 
> other reports of failure relating to compiler warnings.
> All the other platforms passed the newer versions. There is no mention 
> of the UCRT platform at the CRAN Policies page: 
> https://cran.r-project.org/web/packages/policies.html
> I'm waiting to see whether the aging fortran code generates warnings 
> whenever the UCRT platform gets fired up again.

Follow-up on UCRT package checks for muhaz package. When the version 
that corrected R code errors was eventually run, it now has a status of 
"OK". Apparently the warnings about the Fortran code are not 
sufficiently worrisome. It still have an LTO link but the status page 
doesn't even consider the warnings contained therein to be worthy of 
being called warnings or notes.


David Winsemius

More information about the R-package-devel mailing list