Skip to content

[R-pkg-devel] Question regarding CRAN submission with notes

3 messages · Tony Wilkes, Uwe Ligges, Duncan Murdoch

Tony Wilkes
# 
Dear all,

I'm trying to publish an R package to CRAN. Their checks come back with 2 NOTES. The first one is saying that the package is a new submission, and the second one is that there is a subdirectory with a package inside my package. Both notes are correct. I have explained in my submission comments that this is indeed a new submission. I also gave the following explanation for the second note in my submission comments: "There is 1 NOTE given by the R CMD check results. This is due to the fact that I have placed fake libraries with fake packages inside the "inst/tinytest/" folder. This was necessary to test the various import-system related functions in my package. As stated, these are fake packages, and only contain trivial functions of no consequence. These fake libraries with fake packages ONLY exist because it is necessary for proper and extensive unit testing, that is all."

They asked me to fix these notes a few days ago. I replied (politely) repeating my comments a few days ago. I have not heard back from CRAN since. I am assuming that I am missing something, but CRAN won't explain what exactly I am missing.

Here are the NOTES:
Windows: <https://emea01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwin-builder.r-project.org%2Fincoming_pretest%2Ftinycodet_0.1.0.4_20231009_210227%2FWindows%2F00check.log&data=05%7C01%7C%7Ce3fe951ba4c54110abbb08dbc8fb77c2%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C638324754614049791%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=Xh6fZd6iQjW6AOFtCAwwzhNO%2B5vyCh%2Fy5%2FEi25z7D%2Fg%3D&reserved=0<https://win-builder.r-project.org/incoming_pretest/tinycodet_0.1.0.4_20231009_210227/Windows/00check.log>>
Status: 1 NOTE
Debian: <https://emea01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwin-builder.r-project.org%2Fincoming_pretest%2Ftinycodet_0.1.0.4_20231009_210227%2FDebian%2F00check.log&data=05%7C01%7C%7Ce3fe951ba4c54110abbb08dbc8fb77c2%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C638324754614049791%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=0jOFqYf9Ar%2FBp1YVSZjfRHRdChzGPi5lSSuyLT2HAoM%3D&reserved=0<https://win-builder.r-project.org/incoming_pretest/tinycodet_0.1.0.4_20231009_210227/Debian/00check.log>>
Status: 2 NOTEs

More details are given in the directory:
<https://emea01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fwin-builder.r-project.org%2Fincoming_pretest%2Ftinycodet_0.1.0.4_20231009_210227%2F&data=05%7C01%7C%7Ce3fe951ba4c54110abbb08dbc8fb77c2%7C84df9e7fe9f640afb435aaaaaaaaaaaa%7C1%7C0%7C638324754614049791%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=73zTGLwQVRlTMnk20rbgTBOn5dCdu16y%2BmhlLS8GIpg%3D&reserved=0<https://win-builder.r-project.org/incoming_pretest/tinycodet_0.1.0.4_20231009_210227/>>

Could anyone tell me what exactly I'm doing wrong? Thank you in advance.

Kind regards,

Tony
Uwe Ligges
# 
This is under discussion with the CRAN team now.

Best,
Uwe Ligges
On 11.10.2023 09:06, Tony Wilkes wrote:
Duncan Murdoch
# 
The note says it contains "an installed version" of the package, 
presumably referring to your test package.

I think you shouldn't do that.  For a trivial package it might be true 
that the installed version is the same on all systems, but it won't 
continue to be as the installation code evolves.

What you could try is to include files that would let you install the 
test package into a library under tempdir().  Your tests would start by 
installing the test package there, then run the tests, then clean up 
afterwards.

It's possible the tests will still complain about that, in which case 
I'd go one step further back, and write a function to create the source 
package, rather than including it.  For example, have a file 
DESCRIPTION.txt which your test renames to DESCRIPTION, etc.

Duncan Murdoch
On 11/10/2023 3:06 a.m., Tony Wilkes wrote: