// A nil argument is equivalent to an empty slice.
func Compare(a, b []byte) int {
	return bytealg.Compare(a, b)
}

// explode splits s into a slice of UTF-8 sequences, one per Unicode code point (still slices of bytes),
// up to a maximum of n byte slices. Invalid UTF-8 sequences are chopped into individual bytes.
func explode(s []byte, n int) [][]byte {
	if n <= 0 || n > len(s) {
		n = len(s)
	}
	a := make([][]byte, n)

	// in CP-437 (which is mostly compatible with ASCII), unless the UTF-8
	// flag bit is set. However, there are several problems:
	//
	//	* Many ZIP readers still do not support UTF-8.
	//	* If the UTF-8 flag is cleared, several readers simply interpret the
	//	name and comment fields as whatever the local system encoding is.
	//

and less like C and C++, where the meaning of any program with a race
is entirely undefined, and the compiler may do anything at all.
Go's approach aims to make errant programs more reliable and easier to debug,
while still insisting that races are errors and that tools can diagnose and report them.
</p>

<h2 id="model">Memory Model</h2>

<p>
The following formal definition of Go's memory model closely follows