There were no ERRORs or WARNINGs.
There was 1 NOTE associated with a URL that was flagged as invalid.
Maintainer: 'Clay Morrow <[email protected]>'
New submission
Found the following (possibly) invalid URLs:
URL: https://support.posit.co/hc/en-us/articles/200486498-Package-Development-Prerequisites
From: README.md
Status: 403
Message: Forbidden
I checked the URL (on 2023-08-30), which is in the README file, and it works properly.
This version of the package was tested on 2023-08-23, using check-r-package
GitHub Action from the
r-lib repository, with the following call to rcmdcheck
:
rcmdcheck::rcmdcheck(args = c("--no-manual", "--as-cran"), error_on = "error", check_dir = "check")
Checks were conducted on the following systems, using :
-
windows-latest (release)
-
macOS-latest (release)
-
ubuntu-latest (devel)
-
ubuntu-latest (oldrel-1)
-
ubuntu-latest (release)
Additionally, the package was checked on the development version of R for Windows
(on 2023-08-30) with devtools::check_win_devel()
and the same results as above
were returned.
There are currently no downstream dependencies for this package.
-
Some users have had trouble building the vignette on personal machines (using
devtools::install_github()
), but I have not had any trouble on any of the systems I have personally tested with. -
Some examples in
partGLS.Rd
andfitGLS_opt.Rd
can take a bit longer to run (10-20 seconds total) since the functions documented in these files need to invert multiple large matrices.
After initial submission, I received correspondence from Benjamin Altmann at CRAN. In response to the requested changes I have:
- removed trailing white space from all lines in the DESCRIPTION file
- added
@return
values to all exported functions. - removed
:::
from all documentation - removed all
@examples
from un-exported functions - removed examples wrapped in
\dontrun{}
and replaced them with\donttest{}
where necessary. - removed commented out code in examples
Next, it appears that the examples for partGLS
ran on CRAN's check system using more than 2 cores. We were unable
to replicate these results on any of our test machines ((user + system) / elapsed < 2.5
for all test cases) . But,
to be safe we have removed the examples that use multiple cores (ncores=2
) and have set the default condition to use
1 core instead of the C++
compiler's default (for operations using Eigen
). Hopefully this fixes the issue.