Gábor Csárdi

I just renamed `master` to `old`, I am pretty sure that it is mostly defunct now, anyway.

I am sorry for the inconvenience. `@master` was never meant for public use, actually, and many `@v2` actions are much better than `@master`, which is still available under the name...

Yes, I am going to add this: https://github.com/r-lib/rcmdcheck/issues/12 to solve this very issue. It is mostly straightforward, the slightly challenging part is non-deterministic output.

OTOH, maybe we could just suppress particular NOTEs using env vars? What are the ones that we want to suppress? I could think of two: - NOTE about the package...

Or maybe the best is a combination of both: if `NOT_CRAN=true` then we set a list of env vars for the check, and the env vars can be configured in...

> @gaborcsardi see above :) > I think @jimhester's worry is that we'll eventually hit a note that we can't easily suppress, That's true, but CRAN is quite meticulously adding...

Yeah, that's not too bad, either. But if it is in rcmdcheck, then it is used locally as well.

`NOT_CRAN` would just make rcmdcheck add the env vars from `inst/check.env` for the check. There is also another env var to force turn it on/off. So nothing changes, unless people...

@jimhester @hadley https://github.com/r-lib/rcmdcheck/pull/160

Sure, but then it does not work on places where you use a package `.tar.gz`, eg. on R-hub.