Enum BytepackedIntegerEncoder

pub enum BytepackedIntegerEncoder {
    U8(U8BytePacker),
    U16(U16BytePacker),
    U32(U32BytePacker),
    U64(U64BytePacker),
    Zero,
}

A bytepacked integer encoder that automatically chooses the smallest possible integer type to store the given values.

This is byte packing (not bit packing). Not even that, we only fit things into sizes of 1,2,4,8 bytes. It’s simple, fast, and easy but doesn’t provide the maximum possible compression.

Still, it’s useful for things like offsets which are often small and fit into a u16 or u32 but sometimes might need the full u64 range.

In the future we can investigate replacing this with something more sophisticated.

Variants

U8(U8BytePacker)

U16(U16BytePacker)

U32(U32BytePacker)

U64(U64BytePacker)

Zero

Implementations

impl BytepackedIntegerEncoder

pub fn with_capacity(capacity: usize, max_value: u64) -> Self

Create a new encoder with the given capacity and maximum value.

pub unsafe fn append(&mut self, value: u64)

Append a value to the encoder.

Safety

This function is unsafe because it doesn’t check for overflow. If the value is too large to fit in the chosen integer type, it will be silently truncated.

pub fn into_data(self) -> Vec<u8>

Convert the encoder into a vector of bytes.