Skip to content

[R-pkg-devel] Problem with "additional repository".

7 messages · Rolf Turner, Iris Simmons, Duncan Murdoch +2 more

#
I have submitted a new package to CRAN, and this package has been
knocked back on the basis of a NOTE:
This suggested package consists of data sets, the size of which is too
large to satisfy CRAN's constraints. I put this package in a repository
on github, from which it can be accessed by users.

My DESCRIPTION file contains the line:
The given URL seems to work, in that users can indeed load the
ionChannelData package via the command
I was under the impression that this was all that I needed to do.  The
CRAN checking process acknowledges the existence of the repository in
question:
So CRAN knows about this repository.  Why can it not make use of it?

What can/should I do to resolve this problem?

I guess I could simply *not* Suggest ionChannelData.  But what then, is
the point of the option of including an Additional_repositories field in
the DESCRIPTION file?

cheers,

Rolf Turner
#
The packages in additional repositories are not installed on the CRAN
checking machines. The NOTE is expected, but it's not a warning, so your
package should still be allowed to upload to CRAN. You should not remove
ionChannelData from Suggests because then the reference to that package in
your documentation will become a WARNING and your package will not be
uploaded to CRAN.
On Sun, Oct 15, 2023, 21:35 Rolf Turner <rolfturner at posteo.net> wrote:

            

  
  
#
On 15/10/2023 9:34 p.m., Rolf Turner wrote:
Is that web site set up as a repository?  My impression was that it 
needed to be, so people could do

   install.packages("ionChannelData", repos = 
"https://rolfturner.r-universe.dev")

You can use the "drat" package to set up a tiny repository quickly.

Duncan
#
Whoops, I just read the next line.  Sorry!
On 15/10/2023 9:34 p.m., Rolf Turner wrote:
#
Certainly there was more in the output that caused rejection as the 
stuff you describe below seems to be fine.
I do not know which package this refers to, so cannot easily look.

Best,
Uwe
On 16.10.2023 11:25, Duncan Murdoch wrote:
1 day later
#
On Mon, 16 Oct 2023 17:08:28 +0200
Uwe Ligges <ligges at statistik.tu-dortmund.de> wrote:

            
This seems to be about the eglhmm package. It seems to pass --as-cran
checks on my machine, but I know it's not the full set of checks
performed for new packages.
#
It was 0.0-20 that had another issue now explained privately. 0.0-21 
passes cleanly.

Best,
Uwe
On 17.10.2023 20:45, Ivan Krylov wrote: