Post #2352280

2025-02-06 03:29 UTC

Looking into some #rstats code attached to a paper. The code won&#39;t run now, because it relies on a package that has been archived (removed) from CRAN. This package had been around for over a decade, with updates over the years. Looking into why it was removed, it appears that one of the package examples required downloading a GeoTIFF file from a website. It seems that, at the time CRAN tried to run and test this code, the server hosting that file happened to have an expired SSL certificate. The certificate and website is now fine. The example code was marked &quot;don&#39;t run&quot;. Maintainers, understandably, gave up on CRAN&#39;s bullshit and have not bothered with the package since, or trying to get it back on CRAN. Good job, CRAN. https://meteoexploration.com/R/insol/

Replies (4)

@HydrePrever@mathstodon.xyz 2025-02-06 04:30

@jimbob@aus.social the fact that the "don't run" examples are tested never ceases to amaze me. I have had to comment some examples to avoid this. How convenient for the user... Also, checking _R_CHECK_LIMIT_CORES_ on package load to totally disable multi threading was at some point the only way I found to stay on cran EDIT m'y bad, it's the "don't test" examples that are tested... The"don't run" seems safe

Open ##3186966
@barbone@fosstodon.org 2025-02-06 05:04

@jimbob@aus.social While the SSL certificate is unfortunate, the part that failed was not wrapped with `\dontrun{} -- it was the download.file() after it. Oddly there's a second \dontrun{} afterwards. https://cran-archive.r-project.org/web/checks/2023/2023-02-27_check_results_insol.html https://github.com/cran/insol/blob/4fe6f26351af22121a5319b4a8b4950b66ca6226/man/hillshading.Rd#L60-L61

Open ##3186967
@jospueyo@fosstodon.org 2025-02-06 05:59

@jimbob@aus.social CRAN has a lot to improve, I agree with this. However, the fact that CRAN is so restrictive is one of the reasons why installing dependencies in R is usually an easy and smooth process. Comparing, for example, to the Python nightmare. #Rstats

Open ##3186968
@gavin@fosstodon.org 2025-02-06 08:26

@jimbob@aus.social I don't have time to test the use of nix for this, but if it is archived in nix, you could run the {insol} code in a nix-shell, and get around the CRAN issue.

Open ##3186976