webmetro/src/lib.rs

extern crate futures;

#[derive(Debug, PartialEq)]
pub enum Error {
    CorruptVarint,
    UnknownElementId,
    UnknownElementLength,
    CorruptPayload,
}

#[derive(Debug, PartialEq)]
pub enum Varint {
    /// a numeric value
    Value(u64),
    /// the reserved "unknown" value
    Unknown
}

/// Try to parse an EBML varint at the start of the given slice.
/// Returns an Err() if the format is corrupt.
/// Returns Ok(None) if more bytes are needed to get a result.
/// Returns Ok(Some((varint, next))) to return a varint value and
/// the size of the parsed varint.
pub fn decode_varint(bytes: &[u8]) -> Result<Option<(Varint, usize)>, Error> {
    let mut value: u64 = 0;
    let mut value_length = 1;
    let mut mask: u8 = 0x80;
    let mut unknown_marker: u64 = !0;

    if bytes.len() == 0 {
        return Ok(None)
    }

    // get length marker bit from first byte & parse first byte
    while mask > 0 {
        if (mask & bytes[0]) != 0 {
            value = (bytes[0] & !mask) as u64;
            unknown_marker = (mask - 1) as u64;
            break
        }
        value_length += 1;
        mask = mask >> 1;
    }

    if mask == 0 {
        return Err(Error::CorruptVarint)
    }

    // check we have enough data to parse
    if value_length > bytes.len() {
        return Ok(None)
    }

    // decode remaining bytes
    for i in 1..value_length {
        value = (value << 8) + (bytes[i] as u64);
        unknown_marker = (unknown_marker << 8) + 0xFF;
    }

    // determine result
    if value == unknown_marker {
        Ok(Some((Varint::Unknown, value_length)))
    } else {
        Ok(Some((Varint::Value(value), value_length)))
    }
}

/// Try to parse an EBML element header at the start of the given slice.
/// Returns an Err() if the format is corrupt.
/// Returns Ok(None) if more bytes are needed to get a result.
/// Returns Ok(Some((id, varint, next))) to return the element id,
/// the size of the payload, and the size of the parsed varint.
pub fn decode_tag(bytes: &[u8]) -> Result<Option<(u64, Varint, usize)>, Error> {
    // parse element ID
    match decode_varint(bytes) {
        Ok(None) => Ok(None),
        Err(err) => Err(err),
        Ok(Some((Varint::Unknown, _))) => Err(Error::UnknownElementId),
        Ok(Some((Varint::Value(element_id), id_size))) => {
            // parse payload size
            match decode_varint(&bytes[id_size..]) {
                Ok(None) => Ok(None),
                Err(err) => Err(err),
                Ok(Some((element_length, length_size))) =>
                    Ok(Some((
                        element_id,
                        element_length,
                        id_size + length_size
                    )))
            }
        }
    }
}

#[cfg(test)]
mod tests {

    use futures::future::{ok, Future};
    use super::*;
    use super::Error::{CorruptVarint, UnknownElementId};
    use super::Varint::{Unknown, Value};

    #[test]
    fn hello_futures() {
        let my_future = ok::<String, ()>("Hello".into())
            .map(|hello| hello + ", Futures!");

        let string_result = my_future.wait().unwrap();

        assert_eq!(string_result, "Hello, Futures!");
    }

    #[test]
    fn fail_corrupted_varints() {
        assert_eq!(decode_varint(&[0]), Err(CorruptVarint));
        assert_eq!(decode_varint(&[0, 0, 0]), Err(CorruptVarint));
    }

    #[test]
    fn incomplete_varints() {
        assert_eq!(decode_varint(&[]), Ok(None));
        assert_eq!(decode_varint(&[0x40]), Ok(None));
        assert_eq!(decode_varint(&[0x01, 0, 0]), Ok(None));
    }

    #[test]
    fn parse_varints() {
        assert_eq!(decode_varint(&[0xFF]), Ok(Some((Unknown, 1))));
        assert_eq!(decode_varint(&[0x7F, 0xFF]), Ok(Some((Unknown, 2))));
        assert_eq!(decode_varint(&[0x80]), Ok(Some((Value(0), 1))));
        assert_eq!(decode_varint(&[0x81]), Ok(Some((Value(1), 1))));
        assert_eq!(decode_varint(&[0x40, 52]), Ok(Some((Value(52), 2))));

        // test extra data in buffer
        assert_eq!(decode_varint(&[0x83, 0x11]), Ok(Some((Value(3), 1))));
    }

    #[test]
    fn fail_corrupted_tags() {
        assert_eq!(decode_tag(&[0]), Err(CorruptVarint));
        assert_eq!(decode_tag(&[0x80, 0]), Err(CorruptVarint));
        assert_eq!(decode_tag(&[0xFF, 0x80]), Err(UnknownElementId));
        assert_eq!(decode_tag(&[0x7F, 0xFF, 0x40, 0]), Err(UnknownElementId));
    }

    #[test]
    fn incomplete_tags() {
        assert_eq!(decode_tag(&[]), Ok(None));
        assert_eq!(decode_tag(&[0x80]), Ok(None));
        assert_eq!(decode_tag(&[0x40, 0, 0x40]), Ok(None));
    }

    #[test]
    fn parse_tags() {
        assert_eq!(decode_tag(&[0x80, 0x80]), Ok(Some((0, Value(0), 2))));
        assert_eq!(decode_tag(&[0x81, 0x85]), Ok(Some((1, Value(5), 2))));
        assert_eq!(decode_tag(&[0x80, 0xFF]), Ok(Some((0, Unknown, 2))));
        assert_eq!(decode_tag(&[0x80, 0x7F, 0xFF]), Ok(Some((0, Unknown, 3))));
        assert_eq!(decode_tag(&[0x85, 0x40, 52]), Ok(Some((5, Value(52), 3))));
    }
}