numr_replace() or numr_substitute()?

Which to use as the default for formatting numerals? `numr_replace()` uses string replacement; `numr_substitute()` decomposes a numeric into digits and directly subsets the numerals matrix.

I was hoping `numr_substitute()` would be faster, but it isn't (yet):

``` r
library("numerals")

x <- floor(runif(1000) * 1e6)

bench::mark(
numr_substitute(x, "fa")
)
#> Warning: Some expressions had a GC in every iteration; so filtering is disabled.
#> # A tibble: 1 x 6
#> expression min median `itr/sec` mem_alloc `gc/sec`
#>
#> 1 numr_substitute(x, "fa") 18.9ms 21ms 45.2 NA 68.7

bench::mark(
numr_replace(x, "fa")
)
#> # A tibble: 1 x 6
#> expression min median `itr/sec` mem_alloc `gc/sec`
#>
#> 1 numr_replace(x, "fa") 8.51ms 8.54ms 112. NA 0
```

But I do suspect `numr_substitute()` is more robust to e.g. character encoding and locale issues.