Struct BitWriter 

pub struct BitWriter { /* private fields */ }

Writes bits into a growable byte buffer.

Bits are written in little-endian order, with the least significant bit of each value written first. This matches the JXL bitstream format.

Example

use jxl_encoder::bit_writer::BitWriter;

let mut writer = BitWriter::new();
writer.write(8, 0x12).unwrap();
writer.write(4, 0x3).unwrap();
writer.write(4, 0x4).unwrap();
writer.zero_pad_to_byte();

let bytes = writer.finish();
assert_eq!(bytes, vec![0x12, 0x43]);

Implementations

impl BitWriter

pub fn new() -> Self

Creates a new empty BitWriter.

pub fn with_capacity(capacity_bytes: usize) -> Self

Creates a new BitWriter with pre-allocated capacity.

Arguments

• capacity_bytes - Initial capacity in bytes.

pub fn bits_written(&self) -> usize

Returns the total number of bits written.

pub fn bytes_written(&self) -> usize

Returns the number of bytes written (rounded up).

pub fn is_byte_aligned(&self) -> bool

Returns true if the writer is aligned to a byte boundary.

pub fn bits_to_byte_boundary(&self) -> usize

Returns the number of bits needed to reach the next byte boundary.

pub fn write(&mut self, n_bits: usize, bits: u64) -> Result<()>

Writes up to 56 bits to the buffer.

Bits are written in little-endian order, least-significant-bit first. The value must fit in n_bits bits.

Arguments

• n_bits - Number of bits to write (0-56).
• bits - The value to write. Only the lower n_bits are used.

Errors

Returns an error if n_bits > 56 or if allocation fails.

pub fn zero_pad_to_byte(&mut self)

Writes zeros to pad to the next byte boundary.

If already byte-aligned, this is a no-op.

pub fn append_bytes(&mut self, data: &[u8]) -> Result<()>

Appends byte-aligned data from a slice.

The writer must be byte-aligned before calling this method.

Errors

Returns an error if the writer is not byte-aligned.

pub fn append_byte_aligned(&mut self, other: &BitWriter) -> Result<()>

Appends another BitWriter’s contents.

Both writers must be byte-aligned.

Errors

Returns an error if either writer is not byte-aligned.

pub fn append_unaligned(&mut self, other: &BitWriter) -> Result<()>

Appends another BitWriter’s contents, allowing unaligned data.

This is slower than append_byte_aligned but works with any alignment.

pub fn as_bytes(&self) -> &[u8]

Returns a view of the written bytes.

The writer must be byte-aligned.

Panics

Panics if the writer is not byte-aligned.

pub fn peek_bytes(&self) -> &[u8]

Returns a view of the internal storage for debugging.

Unlike as_bytes(), this does not require byte alignment and returns the raw storage including any partial bytes. Useful for debugging bit-level encoding issues.

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

Consumes the writer and returns the written bytes.

The writer must be byte-aligned.

Panics

Panics if the writer is not byte-aligned.

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

Consumes the writer and returns the written bytes, padding if necessary.

Unlike finish, this will zero-pad to byte alignment if needed.

impl BitWriter

pub fn write_bit(&mut self, bit: bool) -> Result<()>

Writes a single bit (0 or 1).

pub fn write_u8(&mut self, value: u8) -> Result<()>

Writes an 8-bit unsigned integer.

pub fn write_u16(&mut self, value: u16) -> Result<()>

Writes a 16-bit unsigned integer in little-endian order.

pub fn write_u32(&mut self, value: u32) -> Result<()>

Writes a 32-bit unsigned integer in little-endian order.

pub fn write_u32_coder( &mut self, value: u32, d0: u32, d1: u32, d2: u32, d3: u32, u_bits: usize, ) -> Result<()>

Writes a U32 value using the JXL variable-length encoding.

The encoding is selector-based:

• 0: value is d0
• 1: value is d1
• 2: value is d2
• 3: u_bits bits follow, value is d3 + read_bits

Arguments

• value - The value to encode.
• d0, d1, d2, d3 - Direct values for selectors 0-2 and offset for selector 3.
• u_bits - Number of bits for the variable portion (selector 3).

pub fn write_enum_default(&mut self, value: u32) -> Result<()>

Writes an enum value using the jxl-rs default u2S encoding. This uses u2S(0, 1, Bits(4)+2, Bits(6)+18):

• selector 0 → value 0
• selector 1 → value 1
• selector 2 → 2 + Bits(4) = values 2-17
• selector 3 → 18 + Bits(6) = values 18-81

pub fn write_u64_coder(&mut self, value: u64) -> Result<()>

Writes a U64 value using the JXL variable-length encoding.

Matches libjxl’s U64Coder::Write (enc_fields.cc:129-166).

Selector encoding:

• 0: value is 0
• 1: 4 bits follow, value is 1 + read_bits (1-16)
• 2: 8 bits follow, value is 17 + read_bits (17-272)
• 3: 12-bit initial group + varint continuation (8-bit groups with stop bit)

Trait Implementations

impl Clone for BitWriter

fn clone(&self) -> BitWriter

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for BitWriter

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Default for BitWriter

fn default() -> Self

Returns the “default value” for a type.