Performance of every(), some(), and none()

[mgirlich]

I recently discovered every(), some() and none() and think they are great as I frequently used a pattern like any(map_lgl(...)). As @DavisVaughan pointed out in this PR they are unfortunately much slower than the map_lgl(...) pattern unless they exit early (benchmark copied from the other PR)

library(purrr)
library(vctrs)

x <- as.list(1:10000)

fn <- function(x) {
  vec_is(x) || is.null(x)
}
bench::mark(
  all(map_lgl(x, fn)),
  every(x, fn)
)
#> # A tibble: 2 × 6
#>   expression               min   median `itr/sec` mem_alloc `gc/sec`
#>   <bch:expr>          <bch:tm> <bch:tm>     <dbl> <bch:byt>    <dbl>
#> 1 all(map_lgl(x, fn))   15.9ms   20.8ms      47.7    3.91MB     30.7
#> 2 every(x, fn)          42.6ms   43.1ms      23.2  103.95KB    104.

bench::mark(
  any(map_lgl(x, vec_is_list)),
  some(x, vec_is_list)
)
#> # A tibble: 2 × 6
#>   expression                        min   median `itr/sec` mem_alloc `gc/sec`
#>   <bch:expr>                   <bch:tm> <bch:tm>     <dbl> <bch:byt>    <dbl>
#> 1 any(map_lgl(x, vec_is_list))   4.63ms   5.55ms     180.    41.31KB     24.0
#> 2 some(x, vec_is_list)          26.32ms  27.23ms      36.4    8.02KB     57.2

It would be great to improve their performance so that they are as fast/faster than map_lgl(...).