2020-06-02 21:10:44 +00:00
|
|
|
package zstd
|
|
|
|
|
|
|
|
import (
|
|
|
|
"encoding/binary"
|
|
|
|
"errors"
|
|
|
|
"fmt"
|
|
|
|
"io"
|
|
|
|
|
|
|
|
"github.com/klauspost/compress/huff0"
|
|
|
|
)
|
|
|
|
|
|
|
|
type dict struct {
|
|
|
|
id uint32
|
|
|
|
|
2020-09-08 12:04:16 +00:00
|
|
|
litEnc *huff0.Scratch
|
2020-06-02 21:10:44 +00:00
|
|
|
llDec, ofDec, mlDec sequenceDec
|
2020-09-08 12:04:16 +00:00
|
|
|
//llEnc, ofEnc, mlEnc []*fseEncoder
|
|
|
|
offsets [3]int
|
|
|
|
content []byte
|
2020-06-02 21:10:44 +00:00
|
|
|
}
|
|
|
|
|
2022-12-11 10:10:51 +00:00
|
|
|
const dictMagic = "\x37\xa4\x30\xec"
|
2020-06-02 21:10:44 +00:00
|
|
|
|
2020-09-08 12:04:16 +00:00
|
|
|
// ID returns the dictionary id or 0 if d is nil.
|
|
|
|
func (d *dict) ID() uint32 {
|
|
|
|
if d == nil {
|
|
|
|
return 0
|
|
|
|
}
|
|
|
|
return d.id
|
|
|
|
}
|
|
|
|
|
|
|
|
// DictContentSize returns the dictionary content size or 0 if d is nil.
|
|
|
|
func (d *dict) DictContentSize() int {
|
|
|
|
if d == nil {
|
|
|
|
return 0
|
|
|
|
}
|
|
|
|
return len(d.content)
|
|
|
|
}
|
|
|
|
|
2020-06-02 21:10:44 +00:00
|
|
|
// Load a dictionary as described in
|
|
|
|
// https://github.com/facebook/zstd/blob/master/doc/zstd_compression_format.md#dictionary-format
|
|
|
|
func loadDict(b []byte) (*dict, error) {
|
|
|
|
// Check static field size.
|
|
|
|
if len(b) <= 8+(3*4) {
|
|
|
|
return nil, io.ErrUnexpectedEOF
|
|
|
|
}
|
|
|
|
d := dict{
|
|
|
|
llDec: sequenceDec{fse: &fseDecoder{}},
|
|
|
|
ofDec: sequenceDec{fse: &fseDecoder{}},
|
|
|
|
mlDec: sequenceDec{fse: &fseDecoder{}},
|
|
|
|
}
|
2022-12-11 10:10:51 +00:00
|
|
|
if string(b[:4]) != dictMagic {
|
2020-06-02 21:10:44 +00:00
|
|
|
return nil, ErrMagicMismatch
|
|
|
|
}
|
|
|
|
d.id = binary.LittleEndian.Uint32(b[4:8])
|
|
|
|
if d.id == 0 {
|
|
|
|
return nil, errors.New("dictionaries cannot have ID 0")
|
|
|
|
}
|
|
|
|
|
|
|
|
// Read literal table
|
|
|
|
var err error
|
2020-09-08 12:04:16 +00:00
|
|
|
d.litEnc, b, err = huff0.ReadTable(b[8:], nil)
|
2020-06-02 21:10:44 +00:00
|
|
|
if err != nil {
|
|
|
|
return nil, err
|
|
|
|
}
|
2020-09-08 12:04:16 +00:00
|
|
|
d.litEnc.Reuse = huff0.ReusePolicyMust
|
2020-06-02 21:10:44 +00:00
|
|
|
|
|
|
|
br := byteReader{
|
|
|
|
b: b,
|
|
|
|
off: 0,
|
|
|
|
}
|
|
|
|
readDec := func(i tableIndex, dec *fseDecoder) error {
|
|
|
|
if err := dec.readNCount(&br, uint16(maxTableSymbol[i])); err != nil {
|
|
|
|
return err
|
|
|
|
}
|
|
|
|
if br.overread() {
|
|
|
|
return io.ErrUnexpectedEOF
|
|
|
|
}
|
|
|
|
err = dec.transform(symbolTableX[i])
|
|
|
|
if err != nil {
|
|
|
|
println("Transform table error:", err)
|
|
|
|
return err
|
|
|
|
}
|
2021-06-08 12:42:57 +00:00
|
|
|
if debugDecoder || debugEncoder {
|
2020-06-02 21:10:44 +00:00
|
|
|
println("Read table ok", "symbolLen:", dec.symbolLen)
|
|
|
|
}
|
|
|
|
// Set decoders as predefined so they aren't reused.
|
|
|
|
dec.preDefined = true
|
|
|
|
return nil
|
|
|
|
}
|
|
|
|
|
|
|
|
if err := readDec(tableOffsets, d.ofDec.fse); err != nil {
|
|
|
|
return nil, err
|
|
|
|
}
|
|
|
|
if err := readDec(tableMatchLengths, d.mlDec.fse); err != nil {
|
|
|
|
return nil, err
|
|
|
|
}
|
|
|
|
if err := readDec(tableLiteralLengths, d.llDec.fse); err != nil {
|
|
|
|
return nil, err
|
|
|
|
}
|
|
|
|
if br.remain() < 12 {
|
|
|
|
return nil, io.ErrUnexpectedEOF
|
|
|
|
}
|
|
|
|
|
|
|
|
d.offsets[0] = int(br.Uint32())
|
|
|
|
br.advance(4)
|
|
|
|
d.offsets[1] = int(br.Uint32())
|
|
|
|
br.advance(4)
|
|
|
|
d.offsets[2] = int(br.Uint32())
|
|
|
|
br.advance(4)
|
|
|
|
if d.offsets[0] <= 0 || d.offsets[1] <= 0 || d.offsets[2] <= 0 {
|
|
|
|
return nil, errors.New("invalid offset in dictionary")
|
|
|
|
}
|
|
|
|
d.content = make([]byte, br.remain())
|
|
|
|
copy(d.content, br.unread())
|
|
|
|
if d.offsets[0] > len(d.content) || d.offsets[1] > len(d.content) || d.offsets[2] > len(d.content) {
|
|
|
|
return nil, fmt.Errorf("initial offset bigger than dictionary content size %d, offsets: %v", len(d.content), d.offsets)
|
|
|
|
}
|
|
|
|
|
|
|
|
return &d, nil
|
|
|
|
}
|