> Turns out that RStudio replaces the install.packages object in the utils
> package.
> Duncan Murdoch
So RStudio unlocks the bindings and alters the exported environment? That seems
like another reason to stick to the terminal interface.
>> Thank you. tools:::.install_packages works.
> I'm glad it works, but it shouldn't be necessary to use (and is not
> part of the API: not documented to keep working this way).
> Best regards,
> Ivan [Krylov]
Thank you for letting me know. I hope I can avoid using private functions in
future. As noted below, my function seems to work now.
> Try setting a breakpoint in system2 before launching your function:
> Best regards,
> Ivan
Now build.package works as written, so there?s nothing to debug. The problem may
have been that this package is so important to me that I put it in .Rprofile.
The package was not installed for the new version of R, so every R session
started with an annoying error message. Presumably a separate session started
with R CMD would just fail without installing the package. That?s no longer a
problem because the package is now installed. However, I don?t know why the
error message wasn?t clearer, and I?m puzzled that I was able to install
roxygen2 & devtools. Thank you everyone, and I?m sorry if I didn?t give the
right information to diagnose the problem faster.
Regards,
Jorgen.
[[alternative HTML version deleted]]