[R-pkg-devel] CRAN submission struggle

[Ivan Krylov]

On Sun, 17 Dec 2023 15:29:34 +0200
Christiaan Pieterse <pietie.cjp.1908 using gmail.com> wrote:

> But, I've uploaded the newly created package as discussed in my first
> email, available at:
> https://github.com/ChristiaanPieterse/iopspackage2.1.0

Are you sure it wouldn't be better to clean up the existing package
instead of creating unrelated forks? If you're afraid of breaking
something that works, do the work on a separate branch
<https://git-scm.com/book/en/v2/Git-Branching-Branches-in-a-Nutshell>
until it both works as well as your current repo currently does *and*
passes R CMD check --as-cran.

> * checking CRAN incoming feasibility ... [25s] NOTE
> Maintainer: 'C.J Pieterse <pietie.cjp.1908 using gmail.com>'
> New submission

"New submission" is to be expected.

> Unknown, possibly misspelled, fields in DESCRIPTION:
>   'Exports'

"Writing R Extensions" doesn't define a field named "Exports". The
exports are declared in the NAMESPACE file. Since you're using
roxygen2, use its @export tag to export your functions and remove the
Exports: field.

> * checking whether package 'iopspackage' can be installed ... [27s]
> WARNING Found the following significant warnings:
>   Warning: package 'Rcpp' was built under R version 4.3.2

This is probably not a problem with your package (but may be a problem
with the way the machine running R CMD check is set up).

> * checking dependencies in R code ... NOTE
> Namespaces in Imports field not imported from:
>   'openxlsx' 'roxygen2' 'tibble'
>   All declared Imports should be used.

> * checking R code for possible problems ... [12s] NOTE
> IOPS: no visible global function definition for 'createWorkbook'
> IOPS: no visible global function definition for 'addWorksheet'
> IOPS: no visible global function definition for 'writeData'
> IOPS: no visible global function definition for 'saveWorkbook'
> Undefined global functions or variables:
>   addWorksheet createWorkbook saveWorkbook writeData

Are you sure you should be importing roxygen2? You only run
roxygenise() before running R CMD build in order to generate
man/*.Rd and NAMESPACE; I don't think it's used after that.

If you don't use functions from tibble, there's no need to import it
or depend on it either. I also don't see you directly using Rcpp, but
there's no warning about it for some reason.

Use the @importFrom tags to import individual functions that you
actually use (i.e. createWorkbook and friends). See
<https://cran.r-project.org/doc/manuals/R-exts.html#Package-namespaces>
for more information on importing.

Also, remove all library() calls from R/iopspackage2.R. Packages live
in namespaces, not in the global environment; your package should rely
upon the dependency information in DESCRIPTION and NAMESPACE (the
latter generated by roxygen2) for its dependencies.

> > data(ExampleTradeData)  
> Warning in data(ExampleTradeData) :
>   data set 'ExampleTradeData' not found

There's no 'data' directory and no file named ExampleTradeData.* in it.
Data for use by the function data() should be prepared as described in
<https://cran.r-project.org/doc/manuals/R-exts.html#Data-in-packages>.
If you want to use files under inst/extdata, you have to read them
manually:

ETD <- read.csv(system.file(
 file.path('extdata','ExampleTradeData.csv'),
 package = 'iopspackage'
))

> * checking for detritus in the temp directory ... NOTE
> Found the following files/directories:
>   'lastMiKTeXException'

Is this on R-hub? This usually happens on R-hub and doesn't indicate a
problem with your package.

> #' temp_dir <- tempdir()
> #' 
> #' # Set the working directory to a temporary directory
> #' setwd(temp_dir)
<...>
> #' # Clean up the temporary directory
> #' unlink(temp_dir, recursive = TRUE)

Please make it a subdirectory of the session temporary directory:

temp_dir <- tempfile()
dir.create(temp_dir)
...

Removing the session temporary directory is not as bad as directly
overwriting user data (it's all going away after the process is shut
down), but quite a lot of parts of R and other packages rely on
tempdir() being a directory that exists, not to mention that there
could be other temporary files in use by other packages.

-- 
Best regards,
Ivan