Trait ebml_iterable::tools::SignedVint

source ·

pub trait SignedVint: Into<i64> + Copy {
    // Provided methods
    fn as_signed_vint(&self) -> Result<Vec<u8>, ToolError> { ... }
    fn as_signed_vint_with_length(
        &self,
        length: usize
    ) -> Result<Vec<u8>, ToolError> { ... }
}

Expand description

Trait to enable easy serialization to a signed vint.

This is only available for types that can be cast as i64. A signed vint can be written as a variable number of bytes just like a regular vint, but the value portion of the vint is expressed in two’s complement notation.

For example, the decimal number “-33” would be written as [0xDF = 1101 1111]. This value is determined by first taking the two’s complement of 33 [0x21 = 0010 0001] but only using the bits available for the vint value. In this case, that is 7 bits (because the vint marker takes up the 8th bit). The two’s complement is [101 1111]. A handy calculator for two’s complement can be found here. Once the two’s complement has been found, simply prepend the vint marker as usual to get [1101 1111 = 0xDF].

Some more examples:

use ebml_iterable::tools::SignedVint;

assert_eq!(vec![0xDF], (-33i64).as_signed_vint().unwrap());
assert_eq!(vec![0x40, 0xC8], (200i64).as_signed_vint().unwrap());
assert_eq!(vec![0x7F, 0x38], (-200i64).as_signed_vint().unwrap());
assert_eq!(vec![0xFF], (-1i64).as_signed_vint().unwrap());