R custom string literal

[simonbyrne]

It might be nice to be able to call R code via a macro string literal, similar to this proposal: JuliaInterop/MATLAB.jl#29.

Of course, one problem would be that using $ for variable substitution would conflict with R lookup usage.

[randy3k]

I believe there is no problem of using $. With this macro,

macro s_mstr(x)
    quote
        rprint($x)
    end
end

there is no problem in extracting the list element by using $:

julia> s"""
       x = list(a=1,b=2)
       y = x$a
       """
[1] 1

PS: unfortunately, r_mstr is used, so we have to think of something different. I vote for s_mstr.

[simonbyrne]

Sorry, what I meant was that it might be difficult to use $ the same way the proposed mat_str macro does, i.e. for copying the corresponding julia variable to R.

[simonster]

You could allow only $ followed by parens, e.g. $(a), unless that's actually valid R syntax in some context. I believe you could backslash-escape other cases of $ before passing to the parser. Alternatively, it may be that valid uses of $ for interpolation are never valid R syntax, so there would be no ambiguity. (I don't know R well enough to say if this is the case.)

The hardest part in JuliaInterop/MATLAB.jl#29 was determining whether a variable/expression is assigned or used. This is easier for MATLAB than other languages because the MATLAB parser sucks and can't itself handle cases that would be hard for us to handle, but it still turned out to be non-trivial. However, if you can actually get the AST from R, this might not be too hard.

[randy3k]

This can also be closed by #94.