Safe URI decoding/encoding

[Serhii Khoma]

was implemented in #20 (comment)

does it make sense to introduce such change?

because right now it just unsafeThrowError

[Nathan Faubion]

I don't really want to add partiality to a printer, especially since the failure mode is so obscure. I'm inclined to just filter out things that don't encode.

[Serhii Khoma]

ok, this is possible

just want to confirm. You are for the 2d option:

1. right now - fails silently with unsafeThrowError
2. we can - fail silently by omitting the URIComponents that cannot be encoded / decoded
3. we can - fail with an error

was

parse :: forall i o. RouteDuplex i o -> String -> Either Parser.RouteError o
parse (RouteDuplex _ dec) = Parser.run dec

-- | Renders a value of type `i` into a String representation of URI path,
-- | query and fragment (hash).
print :: forall i o. RouteDuplex i o -> i -> String
print (RouteDuplex enc _) = Printer.run <<< enc

now

data RouteParseError
  = RouteParseError RouteError
  | DecodeURIError String

parse :: forall i o. RouteDuplex i o -> String -> Either Parser.RouteParseError o
parse (RouteDuplex _ dec) = Parser.run dec

newtype PrintPathError = EncodeURIComponentError String

-- | Renders a value of type `i` into a String representation of URI path,
-- | query and fragment (hash).
print :: forall i o. RouteDuplex i o -> i -> Either PrintPathError String
print (RouteDuplex enc _) = Printer.run <<< enc