Use of an R_str, or similar, macro? about rcall.jl

[nalimilan]

Would an @R macro work? I suspect many R expressions are also valid Julia syntax, and that way one would benefit from syntax highlighting. Though it certainly makes sense to have a non-standard string literal, in particular if one wants to write several lines of R code using """.

[randy3k]

I suspect many R expressions are also valid Julia syntax, and....

I can list a number of exceptions

1. list subset list$abc and list[["abc"]]
2. variable names containing dots, e.g., obj.num1
3. %*%, %%, %in%, etc and user defined operators.
4. assignment <- and <<-

I believe that R_str with """ quote is much flexible than @R macro,

> R"""
f <- function(x){
   x^2
}
f(2)
"""

[simonbyrne]

I actually made a start on this on the rstr branch, currently unfinished. The idea was to support $ substitution whenever the $ was invalid in R, so that something like

R"sum(a$b)"

would be normal R syntax (looking up the b field of a, but

R"sum($b)"

would insert the Julia object b into the expression.

[dmbates]

Any further progress, Simon? I am running a workshop on "Julia for Statisticians" where RCall is extremely useful and would like to give a general prescription of

R"""
# your R code here
"""

[simonbyrne]

Not yet. I'm looking at some other RCall improvements, so I might try to see what I can do in the next few days.

[dmbates]

@simonbyrne Let me know if I can help in some way.

Would it be worthwhile dropping support for v0.3 in addition to other changes? That could simplify the inline documentation and allow for removal of some of the calls to @compat

[simonbyrne]

I was going to do that. I also want to get precompilation working.