cchecksapi
cchecksapi copied to clipboard
sckott
Ability to ignore some NOTEs in badges
@gaborcsardi
I think we should just start easy, and ignore some, e.g. the one about the package size, for a start. If CRAN is OK with the NOTE, then the package author will be OK as well.
okay, makes sense
I was wondering whether it'd make sense to save the CRAN check state of each package right after release, if one assumes that at this point the checks should be ok. It could help identify NOTEs that are not important to put them in a general whitelist, or a list for each package (if we consider the check state of a package right after release is ok, then what the badge should show is whether things got worse?).
By release I mean first release and release of a new version.
Unfortunately "right after release" is a but fuzzy, because some CRAN platform do not run the checks until somewhat later. Also, getting on CRAN does not mean that it will work fine on all CRAN platforms. They only test on two platforms, Windows + R-devel and Debian/gcc + R-devel. You might still get a failure on other Linuxes or Solaris.
I've just noticed the "worst" badge might be more informative than the "summary" badge since it's not binary.