Struct HuffTable 

pub struct HuffTable {
    pub dng_bug16: bool,
    pub has_diffs: bool,
    /* private fields */
}

A canonical Huffman table with fast lookup.

Fields

dng_bug16: bool

Lossless JPEG/DNG quirk: a length-16 symbol means diff = -32768 with no extra bits.

has_diffs: bool

Whether diff bits follow the symbol (false: the symbol value itself is the result).

Implementations

impl HuffTable

pub fn from_counts( counts: &[u8], values: &[u8], has_diffs: bool, ) -> Result<Self>

Build from JPEG-style counts (codes per length 1..=16) and values.

pub fn from_lengths(lengths: &[(u8, u8)], has_diffs: bool) -> Result<Self>

Build from a list of (code_length, symbol) pairs in canonical order (helper for vendors that specify tables as code lengths).

pub fn max_len(&self) -> u32

Maximum code length in this table.

pub fn decode_symbol<B: BitRead>(&self, bits: &mut B) -> Result<u32>

Decode one symbol (no diff handling).

pub fn decode_diff<B: BitRead>(&self, bits: &mut B) -> Result<i32>

Decode a symbol giving a bit length, then read and sign-extend that many diff bits.

pub fn decode_diff_buffered<B: BitRead>( &self, bits: &mut B, err: &mut bool, ) -> i32

Decode one difference assuming at least 27 bits are buffered (the caller refilled, e.g. with BitRead::fill_to): table hits need no refill and no error path. Rare cases (codes longer than the lookup table, 16-bit differences, invalid codes) fall back to the general decoder; a corrupt code sets *err and yields 0 so hot loops avoid per-sample Results.

pub fn read_diff<B: BitRead>(&self, bits: &mut B, len: u32) -> i32

Read the diff for a given length symbol (handles the 16-bit special case).

Trait Implementations

impl Clone for HuffTable

fn clone(&self) -> Self

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Debug for HuffTable

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

Formats the value using the given formatter.