🥮 mooncakes.io
    Sign in

    lzma-mbt

    Independent MoonBit implementation of liblzma core functionality

    lzma
    xz
    compression
    Download zip
    Version
    0.1.0
    License
    Apache-2.0
    Repository
    Last updated
    24 days ago
    Downloads
    404

    No dependencies

    #lzma-mbt

    MoonBit 实现的 LZMA / LZMA2 / XZ 编解码库，目标对齐 liblzma 的外部可观察行为（格式、校验、错误语义），而不是翻译 C 代码。

    模块名：aurora0x27/lzma-mbt。开发约定见 AGENTS.md，分层与接口见 docs/。

    #状态

    • 默认编码：.xz + preset 6 + CRC64
    • 默认解码：Format::Auto（.xz / LZMA_Alone；裸 LZMA2 须显式指定），内存上限 128 MiB（字典，不是明文长度）
    • LZMA1 / LZMA2 / XZ 可 round-trip；.xz 解码支持同一 Stream 内多个 Block（编码仍写单个 Block）；解码侧有嵌入的 liblzma 向量；编码侧由 CI 交给 xz -d / Python lzma 解回
    • Preset.extreme 尚未实现（UnsupportedFeature）；.xz Block Header 过滤器链支持 LZMA2、Delta -> LZMA2 与全部 BCJ（x86/PowerPC/IA-64/ARM/ARM-Thumb/SPARC/ARM64）-> LZMA2；其它未实现的 filter ID（如 RISC-V）仍 UnsupportedFeature
    • 默认拒绝 .xz Footer 之后或裸 LZMA2 结束标记之后的多余字节；DecodeOptions.concatenated=true 时支持连接的多 .xz Stream 与规范 Stream Padding
    • 编码器是贪心匹配，不保证与 xz -6 字节相同

    兼容性矩阵：docs/compatibility.md

    #使用

    import {
      "aurora0x27/lzma-mbt" @lzma,
    }

    let src = b"hello"
    let compressed = @lzma.encode(src[:])
    let roundtrip = @lzma.decode(compressed[:])

    流式（write + finish 保持分块等价；编码器在 Finish 前缓冲，解码器完整流到达后可提前排出）：

    let dec = @lzma.Decoder::new()
    dec.write(chunk0[:])
    dec.write(chunk1[:])
    let roundtrip = dec.finish()

    指定 LZMA_Alone：

    let enc = @lzma.encode(src[:], options={
      ..@lzma.EncodeOptions::default(),
      format: @lzma.Format::Lzma,
    })

    #构建

    需要 MoonBit toolchain（本仓库 CI 使用 chawyehsu/setup-moonup@v1，版本号见仓库根 moonbit-version）。强制工具链版本 moon 0.1.20260904（内置 core 0.10.12+1634b282e，对应 moonup 发行标识 0.10.12+1634b282e），见 docs/research/toolchain.md；用其它版本时不要静默适配，先停下来确认。编码差分还需要 Python 3（标准库 lzma）以及 xz。

    moon check --deny-warn
moon fmt --check
moon test --deny-warn
python3 scripts/diff_encode.py

    LzmaError

    source
    pub suberror LzmaError {
      InvalidInput(String)
      UnexpectedEof
      InvalidConfiguration(String)
      UnsupportedFeature(String)
      MemoryLimit
      DataError
      BufferTooSmall
      FormatError
    } derive(Eq, 
    Debug    )

    Public error type. Error is a MoonBit reserved name. Match on the tag (err is DataError); string payloads are not a stable API.

    Action

    source
    pub(all) enum Action {
      Run
      SyncFlush
      FullFlush
      Finish
    } derive(Eq, 
    Debug    )

    Check

    source
    pub(all) enum Check {
      None
      Crc32
      Crc64
      Sha256
    } derive(Eq, 
    Debug    )

    CodeResult

    source
    pub(all) struct CodeResult {
      consumed : Int
      produced : Int
      status : Status
    } derive(Eq, 
    Debug    )

    DecodeOptions

    source
    pub(all) struct DecodeOptions {
      format : Format
      memlimit : UInt64?
      concatenated : Bool
      ignore_check : Bool
      filters : Array[FilterSpec]
    } derive(Eq, 
    Debug    )

    DecodeOptions::default

    source
    fn DecodeOptions::default() -> DecodeOptions

    DecodeOptions::validate

    source
    fn DecodeOptions::validate(self : DecodeOptions) -> Unit raise LzmaError

    Decoder

    source
    type Decoder

    Decoder::code

    source
    fn Decoder::code(self : Decoder, input : BytesView, output : OutputBuf, action : Action) -> CodeResult raise LzmaError

    Decoder::finish

    source
    fn Decoder::finish(self : Decoder) -> Bytes raise LzmaError

    Decoder::new

    source
    fn Decoder::new(options? : DecodeOptions) -> Decoder raise LzmaError

    Decoder::reset

    source
    fn Decoder::reset(self : Decoder) -> Unit raise LzmaError

    Decoder::total_in

    source
    fn Decoder::total_in(self : Decoder) -> UInt64

    Decoder::total_out

    source
    fn Decoder::total_out(self : Decoder) -> UInt64

    Decoder::write

    source
    fn Decoder::write(self : Decoder, input : BytesView) -> Unit raise LzmaError

    EncodeOptions

    source
    pub(all) struct EncodeOptions {
      format : Format
      preset : Preset
      check : Check
      props : LzmaProps?
      filters : Array[FilterSpec]
    } derive(Eq, 
    Debug    )

    EncodeOptions::default

    source
    fn EncodeOptions::default() -> EncodeOptions

    EncodeOptions::resolved_props

    source
    fn EncodeOptions::resolved_props(self : EncodeOptions) -> LzmaProps raise LzmaError

    EncodeOptions::validate

    source
    fn EncodeOptions::validate(self : EncodeOptions) -> Unit raise LzmaError

    Encoder

    source
    type Encoder

    Encoder::code

    source
    fn Encoder::code(self : Encoder, input : BytesView, output : OutputBuf, action : Action) -> CodeResult raise LzmaError

    Encoder::finish

    source
    fn Encoder::finish(self : Encoder) -> Bytes raise LzmaError

    Encoder::new

    source
    fn Encoder::new(options? : EncodeOptions) -> Encoder raise LzmaError

    Encoder::reset

    source
    fn Encoder::reset(self : Encoder) -> Unit raise LzmaError

    Encoder::total_in

    source
    fn Encoder::total_in(self : Encoder) -> UInt64

    Encoder::total_out

    source
    fn Encoder::total_out(self : Encoder) -> UInt64

    Encoder::write

    source
    fn Encoder::write(self : Encoder, input : BytesView) -> Unit raise LzmaError

    FilterSpec

    source
    pub(all) enum FilterSpec {
      Lzma1(LzmaProps)
      Lzma2(LzmaProps)
      Delta(distance~ : Int)
      BcjX86(start_offset~ : UInt)
    } derive(Eq, 
    Debug    )

    Extra pre/post filters around the codec. Lzma1/Lzma2 here are configuration errors.

    Format

    source
    pub(all) enum Format {
      Auto
      Lzma
      Lzma2
      Xz
    } derive(Eq, 
    Debug    )

    Compressed container / filter format. Auto is decode-only.

    LzmaProps

    source
    pub(all) struct LzmaProps {
      lc : Int
      lp : Int
      pb : Int
      dict_size : UInt
    } derive(Eq, 
    Debug    )

    LZMA1 properties. All fields are required; lc + lp must be at most 4.

    LzmaProps::from_preset

    source
    fn LzmaProps::from_preset(preset : Preset) -> LzmaProps raise LzmaError

    liblzma preset dictionary sizes: 1 << [18,20,21,22,22,23,23,24,25,26].

    LzmaProps::validate

    source
    fn LzmaProps::validate(self : LzmaProps) -> Unit raise LzmaError

    OutputBuf

    source
    pub(all) struct OutputBuf {
      data : Array[Byte]
      start : Int
    }

    Caller-owned output slot. code writes at data[start] and advances start. size is the number of writable bytes (not a hint).

    OutputBuf::new

    source
    fn OutputBuf::new(size? : Int) -> OutputBuf

    OutputBuf::remaining

    source
    fn OutputBuf::remaining(self : OutputBuf) -> Int

    Writable bytes left (data.length() - start). Out-of-range start is 0.

    OutputBuf::written

    source
    fn OutputBuf::written(self : OutputBuf) -> Bytes

    Bytes already written, data[0:start].

    Preset

    source
    pub(all) struct Preset {
      level : Int
      extreme : Bool
    } derive(Eq, 
    Debug    )

    Preset::default

    source
    fn Preset::default() -> Preset

    Preset::new

    source
    fn Preset::new(level : Int, extreme? : Bool) -> Preset raise LzmaError

    Status

    source
    pub(all) enum Status {
      Ok
      StreamEnd
      NeedInput
      NeedOutput
    } derive(Eq, 
    Debug    )

    DEFAULT_MEMLIMIT

    source
    let DEFAULT_MEMLIMIT : UInt64

    Default decoder memory limit when DecodeOptions.memlimit is None.

    MAX_DICT_SIZE

    source
    let MAX_DICT_SIZE : UInt

    MIN_DICT_SIZE

    source
    let MIN_DICT_SIZE : UInt

    UNKNOWN_SIZE

    source
    let UNKNOWN_SIZE : UInt64

    crc32

    source
    fn crc32(data : BytesView, init? : UInt) -> UInt

    crc64

    source
    fn crc64(data : BytesView, init? : UInt64) -> UInt64

    decode

    source
    fn decode(input : BytesView, options? : DecodeOptions) -> Bytes raise LzmaError

    effective_memlimit

    source
    fn effective_memlimit(limit : UInt64?) -> UInt64

    encode

    source
    fn encode(input : BytesView, options? : EncodeOptions) -> Bytes raise LzmaError

    pack_props

    source
    fn pack_props(lc : Int, lp : Int, pb : Int) -> Byte raise LzmaError

    sha256

    source
    fn sha256(data : BytesView) -> Bytes

    unpack_props

    source
    fn unpack_props(props : Byte) -> (Int, Int, Int) raise LzmaError

    Source Files

    Powered by MoonBit

    Site sourceReport issuePackagesBuild queueSkillsStatistics

    © 2026 mooncakes.io