I've run the latest version of survival through the test suites of 486 of the 565 packages
that depend on it (Depends, Imports, LinkingTo, Suggests), and have a couple small issues
that I'm taking to other authors about.
The exercise turned up a half dozen real errors in my package. I plan to submit early
This leads to three comments/questions.
1. Is there a submission site that would check all? There's several that want to use C
libraries that I don't have on my small test machine, or are just too large (rstan is an
entire ecosystem). It could save one submission cycle for the the CRAN maintainers.
2. The update.packages routine was not smart enough to update the packages that have C
code, although every one of those needed to be re-installed in the newest R-devel.
3. This is more of an opinion. One of the failures was due to the fact that one of the
elements in the coxph object's list shifted position. I did address this, but my sympathy
level was low for a developer that counts on absolute order rather than a name. Would you
bend, or tell them to fix their package?