Gábor Csárdi

Oh, yeah, there is a bug here. The `%` is interpreted in `Rprintf()` or `REprintf`. I'll fix that. But that's independent of the `RSTUDIO` env var being unset. That is...

I'll think about how to detect RStudio from C++ code then....

Looks like a bunch of other env vars are there from C as well: ``` RSTUDIO_WINUTILS=C:/Program Files/RStudio/bin/winutils RSTUDIO_PANDOC=C:/Program Files/RStudio/bin/pandoc RMARKDOWN_MATHJAX_PATH=C:/Program Files/RStudio/resources/mathjax-26 RSTUDIO_MSYS_SSH=C:/Program Files/RStudio/bin/msys-ssh-1000-18 ``` This might be an RStudio bug...

Looks like that if you set the variable explicitly with `Sys.setenv(RSTUDIO = "1")` then it will be there in C++ as well?

Maybe, but I don't know how that would be even possible. It is R calling the external code via `.Call`, RStudio has nothing to do with it AFAICT. Anyway, I...

You can skip the build stage, that is usually not a problem. You can also set `R_REMOTES_STANDALONE="true"` (https://github.com/r-lib/remotes#standalone-mode), and then remotes does not use pkgbuild. If it still fails, then...

I think it is fine to have the default still `FALSE`, even if the parent class is cloneable. In the subclass you might add tags that make the class non-cloneable,...

Oh, right, I forgot that. :(

Yeah, that would be reasonable.

Well, I have similar questions as for the sessioninfo issue. What is remotes doing wrong? Why? What do you think it should do instead? Thanks!