use_cpp11, use_cpp14, use_cpp17

[romainfrancois]

I can't never remember what I have to do for variants of C++.
https://cran.r-project.org/doc/manuals/r-release/R-exts.html#Using-C_002b_002b11-code
https://cran.r-project.org/doc/manuals/r-release/R-exts.html#Using-C_002b_002b14-code
https://cran.r-project.org/doc/manuals/r-release/R-exts.html#Using-C_002b_002b17-code

Should I make the relevant use_cpp_ functions ?

[jennybc]

Sounds good to me! cc @hadley @jimhester

[romainfrancois]

might be an argument to use_rcpp or standalone functions, or both.
Using C++ does not necessarily mean using Rcpp.

[hadley]

Random API thoughts:

You don't have to use Rcpp with C++, but it's certainly the most common way, so I wouldn't expend much effort on an API that doesn't use Rcpp (and looking at use_rcpp() there's really not much there).

We probably should extract out a use_src() which would do:

use_directory("src")
use_git_ignore(c("*.o", "*.so", "*.dll"), "src")

We may also extend use_rcpp() to take a name (a la #117).

[romainfrancois]

This is somewhat orthogonal to Rcpp or not Rcpp, and more about when you want to use some C++ standard with or without Rcpp.

[coatless]

This would be an ideal extension candidate for the use_makevars() function sitting in #601. With this being said, the present implementation would require more versatility by parseing the file in a similar way to the DESCRIPTION parser used by use_description() to check for duplicates. One step at a time though...

[jennybc]

I'm putting this in the backlog w.r.t. preparing the release of v1.5.0. If anyone has immediate plans to move on this, let me know.

[coatless]

@jennybc no immediate plan to move on this. I'll likely have cycles in summer/fall to develop a makevars parser package similar to desc.

Might also be worth noting @jeroen's makeconf package does wonders for understanding a user's compile options.