Decoder iterator returns next() when can't decode full object

[tailhook]

I use the following code to unpack a packet from network:

  let mut dec = Decoder::from_reader(io::Cursor::new(buf));
  match dec.decode::<Packet>().next() {...}

And next() returns None when packet was truncated by some reason. Is it works as intended? I would like to have some more explicit error.

I use cbor 0.3.14 if that matters.

Is there a better interface? Should this behavior be documented?

[BurntSushi]

It returns none if there's no more data to decode IIRC. It returns an error if there was a problem decoding.

I'm not sure what specifically is going wrong in your case because I don't see enough details to be able to reproduce your problem. Could you please provide a minimal code sample that I can run that demonstrates your problem? Thanks.

[tailhook]

Okay, the problem is two-fold. Here is the code:

use std::io::Write;
extern crate bytes;
extern crate cbor;
extern crate rustc_serialize;

use cbor::{Encoder, Decoder};

#[derive(RustcDecodable, RustcEncodable, Debug)]
struct Test {
    value: String,
}

fn main() {
    let mut buf = bytes::ByteBuf::mut_with_capacity(16);
    {
        let mut enc = Encoder::from_writer(&mut buf);
        enc.encode(&[Test {
            value: "hello world hello world hello world".to_string()
        }]).unwrap();
    }
    println!("RESULT len: {}, contents: {:?}", buf.bytes().len(), buf.bytes());
    let mut reader = buf.flip();
    let mut dec = Decoder::from_reader(&mut reader);
    println!("Decoded: {:?}", dec.decode::<Test>().next());
}

For small buffer size it returns:

RESULT len: 16, contents: [161, 101, 118, 97, 108, 117, 101, 120, 35, 104, 101, 108, 108, 111, 32, 119]
Decoded: None

When I change CAPACITY to 64 it outputs the following:

RESULT len: 44, contents: [161, 101, 118, 97, 108, 117, 101, 120, 35, 104, 101, 108, 108, 111, 32, 119, 111, 114, 108, 100, 32, 104, 101, 108, 108, 111, 32, 119, 111, 114, 108, 100, 32, 104, 101, 108, 108, 111, 32, 119, 111, 114, 108, 100]
Decoded: Some(Ok(Test { value: "hello world hello world hello world" }))

(If you wonder if bytes::ByteBuf may have issue, the reading part works the same with Cursor, and writer part returns error Err(Error { repr: Custom(Custom { kind: WriteZero, error: StringError("failed to write whole buffer") }) }), when data is truncated)

[BurntSushi]

Nice find! I've fixed this and uploaded a new release of CBOR to crates.io (0.3.16).

[tailhook]

Great! Thanks for quick response.

Still in this example encoding silently truncates value. Should I file a separate issue for this?

[BurntSushi]

The example where you set the capacity to 16? I don't see how that's an issue with cbor. Are you sure that ByteBuf is a growable buffer? I don't see any documentation for the ByteBuf type (and I admit, I don't know what problems the bytes crate is trying to solve), so I'm not sure what the semantics should be.

If you replace your buffer with io::Cursor::new(vec![]) then things work (because Cursor will grow the underlying buffer if possible, which is the case with Vec).

[tailhook]

No, the buffer is intentionally not growable. But when I try to write_all for more that allowed bytes I get the following error:

Err(Error { repr: Custom(Custom { kind: WriteZero, error: StringError("failed to write whole buffer") }) })

Which I expect to be propagated by encoder.