Test and talk about why id is awesome

[DavisVaughan]

The id argument in combination with the msg %<~% delayed assign operator used internally is awesome. It lets us totally avoid lifecycle_message() in deprecate_soft() and deprecate_warn(), which are expensive.

This PR documents this usage of id, and adds a test so we hopefully don't regress on this in the future. I also added a comment in 26fa727 about this.

# Without `id` (CRAN vs this PR with all the other fixes so far)

cross::bench_versions(pkgs = c("lifecycle", "r-lib/lifecycle#197"), \() {
  library(lifecycle)
  
  # trigger the 8 hour warning once
  deprecate_soft("1.1.0", I("my thing"), details = "for this")

  bench::mark(
    deprecate_soft("1.1.0", I("my thing"), details = "for this"),
    iterations = 100000
  )
})
#> # A tibble: 2 × 7
#>   pkg                 expression                      min  median `itr/sec` mem_alloc `gc/sec`
#>   <chr>               <bch:expr>                    <bch> <bch:t>     <dbl> <bch:byt>    <dbl>
#> 1 lifecycle           "deprecate_soft(\"1.1.0\", I… 224µs 239.8µs     4046.    8.55KB     42.0
#> 2 r-lib/lifecycle#197 "deprecate_soft(\"1.1.0\", I…  43µs  44.4µs    21872.    8.56KB     59.7

# With `id` (CRAN vs this PR with all the other fixes so far)

cross::bench_versions(pkgs = c("lifecycle", "r-lib/lifecycle#197"), \() {
  library(lifecycle)
  
  # trigger the 8 hour warning once
  deprecate_soft("1.1.0", I("my thing"), details = "for this", id = "foo")

  bench::mark(
    deprecate_soft("1.1.0", I("my thing"), details = "for this", id = "foo"),
    iterations = 100000
  )
})
#> # A tibble: 2 × 7
#>   pkg                 expression                      min  median `itr/sec` mem_alloc `gc/sec`
#>   <chr>               <bch:expr>                  <bch:t> <bch:t>     <dbl> <bch:byt>    <dbl>
#> 1 lifecycle           "deprecate_soft(\"1.1.0\",…   145µs 154.9µs     6316.    8.55KB     44.1
#> 2 r-lib/lifecycle#197 "deprecate_soft(\"1.1.0\",…  31.7µs  33.1µs    29172.    8.56KB     67.0

When a once per session message has already been thrown and id is supplied, we are getting very close to "just early exiting" - almost all of the time is now in signalCondition() due to it not being as lazy as we'd hope.

profvis::profvis(for (i in 1:100000) deprecate_soft("1.1.0", I("my thing"), details = "for this", id = "foo"))