Add `encode_vec(input: T, buf: &mut Vec<u8>)` that appends to a `Vec` instead of a `String`

[Not-A-Normal-Robot]

Before you file an issue

• Did you read the docs? yes, and it only has encode_string,
• Did you read the README? yes, and it doesn't seem to have a mention about this

The problem

Let's say I want to write using a specific file format that has base64 interspersed with non-utf8 bytes:

......binary stuff......dmFsaWQgYmFzZTY0Cg==......more binary stuff......

Currently this is a bit tricky to do efficiently. None of the current Engine methods have a thing that supports outputting into a &mut Vec<u8>:

• encode(T) -> String allocates a whole new string, and I'd have to copy it over to my buffer.
• encode_string(T, &mut String) takes in a &mut String so it may not have to reallocate... but I don't have a String here; I only have a Vec<u8> buffer.
• encode_slice(T, &mut [u8]) could kinda work? I'd have to extend my Vec<u8> with zeroes, then slice into those zeroes, making sure it has a sufficient size or else it'll return an error.

It would be a lot more convenient if there were an encode_vec(T, &mut Vec<u8>) method. Not only would this skip zeroing out the memory, it also means unlike encode_slice there's not really a reason for it to fail (besides allocator failures but ehhhh)

How I, the issue filer, am going to help solve it

By proposing this function to be added.

Not a Duplicate (probably)

I actually saw issue #257 about adding encode_base64(input: &[u8]) -> Vec<u8>. This allocates a new Vec. Meanwhile I'm actually suggesting a way to append to an existing Vec. So I don't think this is a duplicate. I didn't really find much else about encoding to Vecs when searching for "Vec" in the issue tracker, but maybe I missed something?

[marshallpierce]

It looks like this functionality is basically a special case of EncoderWriter, as that can work on any Write impl. Perhaps it makes sense to add to Engine, but then again, we wouldn't want to try to cover every conceivable form of buffer-ish container...

[marshallpierce]

Oh, and also:

encode_slice(T, &mut [u8]) could kinda work? I'd have to extend my Vec with zeroes, then slice into those zeroes

This is what an encode_vec impl would do. Without resorting to unsafe, you do have to append zeros, then slice. The only thing that inclusion in this crate would do is we could spare the user doing the length calculation, but it'd still be the same logic.