On 9 August 2017 at 08:38, Duncan Murdoch wrote:
| Nothing that you've written is reproducible by others. See if you can
| put together a short self-contained series of instructions that display
| the problems (or just one of them). It might involve installing from
| Github, or making a package tarball available somewhere.
I am fairly certain I have observed this myself, and I *think* it simply
stems from running R CMD INSTALL (and alike) from a source directory, as
opposed to a tarball. I filed it under minor nuisance and didn't care too
much either as the official two-step of 'R CMD build ...; R CMD INSTALL ...'
is just that: the official and recommended way. With which things work.
http://dirk.eddelbuettel.com | @eddelbuettel | e...@debian.org
Rfirstname.lastname@example.org mailing list