So, there were some edge cases reported that were not handled well by my robotstxt package. I fixed them and did some functionally extension - more robustness here, a little complaining about un-parable-formats here and a new function parameter there …
Now whats the Ropensci procedure for getting the changes on CRAN. I know how to update and puplish packages on CRAN but is there some Ropensci-things I have to consider first or should I just go forward?
You should just go forward. At the moment, we don’t have specific policies related to package updates. We do daily testing of all onboarded packages as they change and we monitor the CRAN status of those on CRAN, and will step in if something breaks and the author doesn’t fix it in a timely way.
An update policy is something we know we have to tackle and hope to get to it in the coming year. At minimum we should be tagging packages to make clear the peer-reviewed version. We have done re-reviews at author request for major updates, but this is probably more work than feasible. It might be good, though, to have “Ropensci-ish things” that we ask authors to consider and editors to monitor.
Good question, I do not think I have thoughts on a general procedure, but:
A lot of work has gone into reviewing, streamlining and improving the package during the onboarding. It would seem only consequent to have some procedure for updates as well. For my particular case these are some things that came up:
1.) What is the procedure - onboarding has one so should updating.
2.) If nobody is watching one might get sloppy on documentation - standards that were required for onboarding might erode over time - we are all busy - right? (was News updated, are there examples for all main functions, are vignettes still working / up-to-date, …)
3.) Implementation details: Although, updates usually do not change that much there might still be little questions some room for discussion, decisions or advise.