Function calcBlockSize

s2/encode_go.go:331–514

input must be > inputMargin

(src []byte, _ *[32768]byte)

// input must be > inputMargin
func calcBlockSize(src []byte, _ *[32768]byte) (d int) {
	// Initialize the hash table.
	const (
		tableBits    = 13
		maxTableSize = 1 << tableBits
	)

	var table [maxTableSize]uint32

	// sLimit is when to stop looking for offset/length copies. The inputMargin
	// lets us use a fast path for emitLiteral in the main loop, while we are
	// looking for copies.
	sLimit := len(src) - inputMargin

	// Bail if we can't compress to at least this.
	dstLimit := len(src) - len(src)>>5 - 5

	// nextEmit is where in src the next emitLiteral should start from.
	nextEmit := 0

	// The encoded form must start with a literal, as there are no previous
	// bytes to copy, so we start looking for hash matches at s == 1.
	s := 1
	cv := load64(src, s)

	// We search for a repeat at -1, but don't output repeats when nextEmit == 0
	repeat := 1

	for {
		candidate := 0
		for {
			// Next src position to check
			nextS := s + (s-nextEmit)>>6 + 4
			if nextS > sLimit {
				goto emitRemainder
			}
			hash0 := hash6(cv, tableBits)
			hash1 := hash6(cv>>8, tableBits)
			candidate = int(table[hash0])
			candidate2 := int(table[hash1])
			table[hash0] = uint32(s)
			table[hash1] = uint32(s + 1)
			hash2 := hash6(cv>>16, tableBits)

			// Check repeat at offset checkRep.
			const checkRep = 1
			if uint32(cv>>(checkRep*8)) == load32(src, s-repeat+checkRep) {
				base := s + checkRep
				// Extend back
				for i := base - repeat; base > nextEmit && i > 0 && src[i-1] == src[base-1]; {
					i--
					base--
				}
				d += emitLiteralSize(src[nextEmit:base])

				// Extend forward
				candidate := s - repeat + 4 + checkRep
				s += 4 + checkRep