std::string_view support in Rcpp::wrap()

[evalon32]

Currently, Rcpp::wrap() treats a std::string_view argument by default as a generic container and creates a character vector from single characters. This is different from how std::string is handled. This change aligns the two, so that wrap("foo"sv) is equivalent to wrap("foo"s). The added code is conditionally compiled for C++17.

Checklist

• Code compiles correctly
• R CMD check still passes all tests
• Preferably, new tests were added which fail without the change
• Document the changes by file in ChangeLog

[eddelbuettel]

Would you mind filling in the template you carried over unchanged?

[evalon32]

This is a draft I created solely to reference in an issue I'm about to open to start a discussion :)

[eddelbuettel]

Looking forward to it -- ex ante it's a good topic and something that would be nice to fill in.

[eddelbuettel]

(Updated your first box above: if you make it [x] without the space you had [ x] then the GH flavored markdown takes it as a completed tick box.)

This looks good. I will toss this at the reverse dependency checker which, being underpowered on an ancient old box, will take a while. We will know more in a few days.

[eddelbuettel]

There were no regressions, or, in the parlance, "changes to worse" so no objections from the reverse-depends check.

[eddelbuettel]

LGTM so merged.