Skip to content

[R-pkg-devel] Discrepancy between R CMD check results and usethis::use_cran_comments

3 messages · Leonard Mada, Ivan Krylov, Hadley Wickham

#
Dear List-Members,

There are no errors/warnings/notes when I run the check:

?? R CMD check results 
???????????????????????????????????????????????????????? Rpdb 2.3.3 ????
Duration: 2m 50.1s

0 errors ? | 0 warnings ? | 0 notes ?

However, there is a discrepancy when I run:
usethis::use_cran_comments(open = rlang::is_interactive())
=>
0 errors ? | 0 warnings ? | 1 note

For some reason, the file is saved with 1 note. The discrepancy remains 
even if I restart R, delete the old cran-comments.md file, and re-run 
the check.


By the way: what is the purpose of this cran-comments.md file? It does 
not seem to be included in the package archive?


Sincerely,


Leonard
#
? Tue, 26 Sep 2023 10:01:21 +0300
Leonard Mada via R-package-devel <r-package-devel at r-project.org> ?????:
Is it R CMD check or R CMD check --as-cran? You will most likely get a
NOTE when checking a package with --as-cran that had been archived on
CRAN.
This convention is from the book "R Packages" by Hadley Wickham and
Jennifer Bryan:
https://r-pkgs.org/release.html#sec-release-cran-comments

According to the book, it's important to put the CRAN submission
comments under version control. Indeed, the actual file is excluded
from the source package using .Rbuildignore. Its contents are
transferred to the "optional comment" field when the devtools
package submits the form at <https://cran.r-project.org/submit.html>.

Many people find the devtools abstraction layer convenient, and since
you're already using devtools, the path of least resistance is to keep
using this file.
#
On Tue, Sep 26, 2023 at 2:01?AM Leonard Mada via R-package-devel
<r-package-devel at r-project.org> wrote:
This is just a template; you need to fill it out with the actual
results. (There's usually one note on first submission which is why we
use this as the default.)

Hadley