Double dispatch doesn't seem to use "the first method" when the methods are different

[yutannihilation]

"14.7.5 Double dispatch" says:

The methods are different, and R calls the first method with a warning.

But the actual implementation seems different; C function DispatchGroup is

    if( lsxp != rsxp ) {
	if ( isFunction(lsxp) && isFunction(rsxp) ) {
	    /* special-case some methods involving difftime */

	    ...snip...

	    else {
		warning(_("Incompatible methods (\"%s\", \"%s\") for \"%s\""),
			lname, rname, generic);
		UNPROTECT(4);
		return 0;
	    }
	}

https://github.com/wch/r-source/blob/fabab8108783c8c48ae277120cfa0046b7b48f17/src/main/eval.c#L3623-L3641

This means do_arith skips this if branch and uses the internal method dispatch, which follows after:

	if (DispatchGroup("Ops", call, op, args, env, &ans))
	    return ans;

https://github.com/wch/r-source/blob/5a156a0865362bb8381dcd69ac335f5174a4f60c/src/main/arithmetic.c#L403-L404

So, I feel the line should be

The methods are different, and R uses the internal dispatch with a warning.

(FWIW, I noticed this when I investigated the error below, which I don't come up with the nice way to do (other than using S4) yet...)

library(ggplot2)

set.seed(100)
d1 <- data.frame(x = 1:100, y = cumsum(runif(100)))
d2 <- data.frame(x = 1:100, y = cumsum(runif(100)))

plot_all <- function(...) {
  l <- lapply(list(...), function(d) ggplot(d, aes(x, y)) + geom_line())
  l <- unname(l)
  class(l) <- "manyplot"
  l
}

print.manyplot <- function(x, ...) {
  do.call(gridExtra::grid.arrange, x)
}

p <- plot_all(d1, d2)
p

`+.manyplot` <- function(e1, e2) {
  l <- lapply(e1, function(x) x + e2)
  class(l) <- "manyplot"
  l
}

p + theme_bw()
#> Warning: Incompatible methods ("+.manyplot", "+.gg") for "+"
#> Error in p + theme_bw(): non-numeric argument to binary operator

[hadley]

Oh good point. Do you want to do a PR?

ggplot2 might eventually end up using the double-dispatch approach in vctrs, once I've worked it out)

[yutannihilation]

Thanks! I'll create a PR.

ggplot2 might eventually end up using the double-dispatch approach in vctrs

Wow, cool!