Remove check for primitive/builtin in as_mapper.default() to improve performance

[mtcarsalot]

• change test to reflect
• The func description did not mention primitives, so i did not explain that these are handled differently now...

Fixes #1088

performance before and after fix.

x <- as.list(1:20)
x[21] <- list(NULL)

bench::mark(
   wrapped = purrr::detect_index(x, function(x) is.null(x)),
   prim = purrr::detect_index(x, is.null)
)
# A tibble: 2 × 13
  expression      min   median `itr/sec` mem_alloc `gc/sec` n_itr  n_gc total_time result    memory              time       gc      
  <bch:expr> <bch:tm> <bch:tm>     <dbl> <bch:byt>    <dbl> <int> <dbl>   <bch:tm> <list>    <list>              <list>     <list>  
1 wrapped      62.3µs   65.6µs    13760.    4.12KB     6.29  6563     3      477ms <int [1]> <Rprofmem [12 × 3]> <bench_tm> <tibble>
2 prim        137.4µs  143.7µs     6502.        0B     6.28  3108     3      478ms <int [1]> <Rprofmem [0 × 3]>  <bench_tm> <tibble>

 devtools::load_all()
ℹ Loading purrr

bench::mark(
   wrapped = purrr::detect_index(x, function(x) is.null(x)),
   prim = purrr::detect_index(x, is.null) 
)
# A tibble: 2 × 13
  expression      min   median `itr/sec` mem_alloc `gc/sec` n_itr  n_gc total_time result    memory              time       gc      
  <bch:expr> <bch:tm> <bch:tm>     <dbl> <bch:byt>    <dbl> <int> <dbl>   <bch:tm> <list>    <list>              <list>     <list>  
1 wrapped        75µs   78.1µs    12529.    4.12KB     6.31  5955     3      475ms <int [1]> <Rprofmem [12 × 3]> <bench_tm> <tibble>
2 prim         69.3µs   72.7µs    13506.    30.8KB     8.52  6338     4      469ms <int [1]> <Rprofmem [68 × 3]> <bench_tm> <tibble>

[hadley]

Could you please add a news bullet too?

[shikokuchuo]

Thanks for your contribution @mtcarsalot! I just made a small edit to the news item and added a comment to the test to make it clear that the current behaviour is intentional.