Remove dependencies

[etiennebacher]

Close #1. Ready for review @JosiahParry ;)

[JosiahParry]

Currently running into a bug that I'm going to think through how we can address it. !b.inherits("raw") does work in our care right now.

First, the check for the raw class isn't working and two, we're returning an Rstr::na() (don't get me started on the oddities of character vectors and the C api) which, instead, we should be returning a NULL

library(b64)
x <- encode(c("hello", "world", NA))

y <- lapply(x, charToRaw)

decode(y)
#> <blob[3]>
#> [[1]]
#> <CHARSXP: NA>
#> 
#> [[2]]
#> <CHARSXP: NA>
#> 
#> [[3]]
#> <CHARSXP: NA>

[JosiahParry]

We can replace the body of the closure with a try_from() coupled with a match statement here.

  let raw = Raw::try_from(b);
  match raw {
      Ok(r) => {
          let decoded = eng.decode(r.as_slice());
          match decoded {
              Ok(d) => Raw::from_bytes(&d).into_robj(),
              Err(_) => ().into_robj(),
          }
      },
      Err(_) => ().into_robj(),
  }

library(b64)
x <- encode(c("hello", "world", NA))

y <- lapply(x, charToRaw)

decode(y)
#> <blob[3]>
#> [[1]]
#> [1] 68 65 6c 6c 6f
#> 
#> [[2]]
#> [1] 77 6f 72 6c 64
#> 
#> [[3]]
#> NULL

[JosiahParry]

Once we fix the vectorized decoding of raws I think this can get merged! Thank you!

[etiennebacher]

Thanks for your help @JosiahParry, is this what you had in mind?

library(b64)
x <- encode(c("hello", "world", NA))
x
#> [1] "aGVsbG8=" "d29ybGQ=" NA

y <- lapply(x, charToRaw)
decode(y)
#> <blob[3]>
#> [[1]]
#> [1] 68 65 6c 6c 6f
#> 
#> [[2]]
#> [1] 77 6f 72 6c 64
#> 
#> [[3]]
#> NULL

# for pretty printing
library(blob)
decode(y)
#> <blob[3]>
#> [1] blob[5 B] blob[5 B] <NA>