{"fofofofofoofofoffofoobarfoo", "foobars", -1},
	{"foofyfoobarfoobar", "y", 4},
	{"oooooooooooooooooooooo", "r", -1},
	{"oxoxoxoxoxoxoxoxoxoxoxoy", "oy", 22},
	{"oxoxoxoxoxoxoxoxoxoxoxox", "oy", -1},
	// test fallback to Rabin-Karp.
	{"000000000000000000000000000000000000000000000000000000000000000000000001", "0000000000000000000000000000000000000000000000000000000000000000001", 5},
}

var lastIndexTests = []BinOpTest{
	{"", "", 0},

		}
		if s[i+1] == c1 && Equal(s[i:i+n], sep) {
			return i
		}
		i++
		fails++
		if fails >= 4+i>>4 && i < t {
			// Give up on IndexByte, it isn't skipping ahead
			// far enough to be better than Rabin-Karp.
			// Experiments (using IndexPeriodic) suggest
			// the cutover is about 16 byte skips.
			// TODO: if large prefixes of sep are matching
			// we should cutover at even larger average skips,

*p /= 2
*p += i
</pre>

<p>
If <code>i</code> and <code>*p</code> start equal to 2,
the original code does <code>*p = 3</code>,
so a racing thread can read only 2 or 3 from <code>*p</code>.
The rewritten code does <code>*p = 1</code> and then <code>*p = 3</code>,
allowing a racing thread to read 1 as well.
</p>

<p>
Note that all these optimizations are permitted in C/C++ compilers: