Expose skip of inst/doc overwrite check directly to rcmdcheck()

[MichaelChirico]

Follow-up to #130 / #155.

AFAICT that solution is to add a new entry to the package DESCRIPTION telling it "it's OK to overwrite inst/doc".

However, I am working on a use case where I clone the package from the GitHub mirror (e.g. github.com/cran/xts) and then check it. The GitHub mirror contains the unpacked tar.gz version of the packages, which always has the inst/doc folder (if there are vignettes). Since I'm just running R CMD check on a git clone & then deleting the package, I definitely don't care if inst/doc gets deleted.

Essentially, I am not the package maintainer, so editing the package DESCRIPTION is a very obscure way to go about this.

For now, I will just unlink('inst/doc', recursive=TRUE) to my clone, but IMHO it would be a better design to expose some way to say "yes" to this prompt directly in rcmdcheck().

[gaborcsardi]

I think there are some issues in pkgbuild that would address this.