diff --git a/src/lib/bytes/Makefile b/src/lib/bytes/Makefile
new file mode 100644
index 0000000000000..7f10585c852b5
--- /dev/null
+++ b/src/lib/bytes/Makefile
@@ -0,0 +1,68 @@
+# Copyright 2009 The Go Authors. All rights reserved.
+# Use of this source code is governed by a BSD-style
+# license that can be found in the LICENSE file.
+
+# DO NOT EDIT.  Automatically generated by gobuild.
+# gobuild -m >Makefile
+
+D=
+
+O_arm=5
+O_amd64=6
+O_386=8
+OS=568vq
+
+O=$(O_$(GOARCH))
+GC=$(O)g -I_obj
+CC=$(O)c -FVw
+AS=$(O)a
+AR=6ar
+
+default: packages
+
+clean:
+	rm -rf *.[$(OS)] *.a [$(OS)].out _obj
+
+test: packages
+	gotest
+
+coverage: packages
+	gotest
+	6cov -g `pwd` | grep -v '_test\.go:'
+
+%.$O: %.go
+	$(GC) $*.go
+
+%.$O: %.c
+	$(CC) $*.c
+
+%.$O: %.s
+	$(AS) $*.s
+
+O1=\
+	bytes.$O\
+
+
+phases: a1
+_obj$D/bytes.a: phases
+
+a1: $(O1)
+	$(AR) grc _obj$D/bytes.a bytes.$O
+	rm -f $(O1)
+
+
+newpkg: clean
+	mkdir -p _obj$D
+	$(AR) grc _obj$D/bytes.a
+
+$(O1): newpkg
+$(O2): a1
+
+nuke: clean
+	rm -f $(GOROOT)/pkg$D/bytes.a
+
+packages: _obj$D/bytes.a
+
+install: packages
+	test -d $(GOROOT)/pkg && mkdir -p $(GOROOT)/pkg$D
+	cp _obj$D/bytes.a $(GOROOT)/pkg$D/bytes.a
diff --git a/src/lib/bytes/bytes.go b/src/lib/bytes/bytes.go
new file mode 100644
index 0000000000000..fe97b04958446
--- /dev/null
+++ b/src/lib/bytes/bytes.go
@@ -0,0 +1,154 @@
+// Copyright 2009 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+// A package of simple functions to manipulate arrays of bytes.
+// Analagous to the facilities of the strings package.
+package bytes
+
+import "utf8"
+
+// Compare returns an integer comparing the two byte arrays lexicographically.
+// The result will be 0 if a==b, -1 if a < b, and +1 if a > b
+func Compare(a, b []byte) int {
+	for i := 0; i < len(a) && i < len(b); i++ {
+		switch {
+		case a[i] > b[i]:
+			return 1
+		case a[i] < b[i]:
+			return -1
+		}
+	}
+	switch {
+	case len(a) < len(b):
+		return -1
+	case len(a) > len(b):
+		return 1
+	}
+	return 0
+}
+
+// Equal returns a boolean reporting whether a == b.
+func Equal(a, b []byte) bool {
+	if len(a) != len(b) {
+		return false
+	}
+	for i := 0; i < len(a); i++ {
+		if a[i] != b[i] {
+			return false
+		}
+	}
+	return true
+}
+
+// Explode splits s into an array of UTF-8 sequences, one per Unicode character (still arrays of bytes).
+// Invalid UTF-8 sequences become correct encodings of U+FFF8.
+func Explode(s []byte) [][]byte {
+	a := make([][]byte, utf8.RuneCount(s));
+	var size, rune int;
+	i := 0;
+	for len(s) > 0 {
+		rune, size = utf8.DecodeRune(s);
+		a[i] = s[0:size];
+		s = s[size:len(s)];
+		i++;
+	}
+	return a
+}
+
+// Count counts the number of non-overlapping instances of sep in s.
+func Count(s, sep []byte) int {
+	if len(sep) == 0 {
+		return utf8.RuneCount(s)+1
+	}
+	c := sep[0];
+	n := 0;
+	for i := 0; i+len(sep) <= len(s); i++ {
+		if s[i] == c && (len(sep) == 1 || Equal(s[i:i+len(sep)], sep)) {
+			n++;
+			i += len(sep)-1
+		}
+	}
+	return n
+}
+
+// Index returns the index of the first instance of sep in s, or -1 if sep is not present in s.
+func Index(s, sep []byte) int {
+	n := len(sep);
+	if n == 0 {
+		return 0
+	}
+	c := sep[0];
+	for i := 0; i+n <= len(s); i++ {
+		if s[i] == c && (n == 1 || Equal(s[i:i+n], sep)) {
+			return i
+		}
+	}
+	return -1
+}
+
+// Split returns the array representing the subarrays of s separated by sep. Adjacent
+// occurrences of sep produce empty subarrays.  If sep is empty, it is the same as Explode.
+func Split(s, sep []byte) [][]byte {
+	if len(sep) == 0 {
+		return Explode(s)
+	}
+	c := sep[0];
+	start := 0;
+	n := Count(s, sep)+1;
+	a := make([][]byte, n);
+	na := 0;
+	for i := 0; i+len(sep) <= len(s); i++ {
+		if s[i] == c && (len(sep) == 1 || Equal(s[i:i+len(sep)], sep)) {
+			a[na] = s[start:i];
+			na++;
+			start = i+len(sep);
+			i += len(sep)-1
+		}
+	}
+	a[na] = s[start:len(s)];
+	return a
+}
+
+// Join concatenates the elements of a to create a single byte array.   The separator
+// sep is placed between elements in the resulting array.
+func Join(a [][]byte, sep []byte) []byte {
+	if len(a) == 0 {
+		return []byte{}
+	}
+	if len(a) == 1 {
+		return a[0]
+	}
+	n := len(sep) * (len(a)-1);
+	for i := 0; i < len(a); i++ {
+		n += len(a[i])
+	}
+
+	b := make([]byte, n);
+	bp := 0;
+	for i := 0; i < len(a); i++ {
+		s := a[i];
+		for j := 0; j < len(s); j++ {
+			b[bp] = s[j];
+			bp++
+		}
+		if i + 1 < len(a) {
+			s = sep;
+			for j := 0; j < len(s); j++ {
+				b[bp] = s[j];
+				bp++
+			}
+		}
+	}
+	return b
+}
+
+// HasPrefix tests whether the byte array s begins with prefix.
+func HasPrefix(s, prefix []byte) bool {
+	return len(s) >= len(prefix) && Equal(s[0:len(prefix)], prefix)
+}
+
+// HasSuffix tests whether the byte array s ends with suffix.
+func HasSuffix(s, suffix []byte) bool {
+	return len(s) >= len(suffix) && Equal(s[len(s)-len(suffix):len(s)], suffix)
+}
diff --git a/src/lib/bytes/bytes_test.go b/src/lib/bytes/bytes_test.go
new file mode 100644
index 0000000000000..26b3fc21d0a0a
--- /dev/null
+++ b/src/lib/bytes/bytes_test.go
@@ -0,0 +1,130 @@
+// Copyright 2009 The Go Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style
+// license that can be found in the LICENSE file.
+
+package bytes
+
+import (
+	"bytes";
+	"io";
+	"testing";
+)
+
+func eq(a, b []string) bool {
+	if len(a) != len(b) {
+		return false;
+	}
+	for i := 0; i < len(a); i++ {
+		if a[i] != b[i] {
+			return false;
+		}
+	}
+	return true;
+}
+
+func arrayOfString(a [][]byte) []string {
+	result := make([]string, len(a));
+	for j := 0; j < len(a); j++ {
+		result[j] = string(a[j])
+	}
+	return result
+}
+
+// For ease of reading, the test cases use strings that are converted to byte
+// arrays before invoking the functions.
+
+var abcd = "abcd"
+var faces = "☺☻☹"
+var commas = "1,2,3,4"
+var dots = "1....2....3....4"
+
+type CompareTest struct {
+	a string;
+	b string;
+	cmp int;
+}
+var comparetests = []CompareTest {
+	CompareTest{ "", "", 0 },
+	CompareTest{ "a", "", 1 },
+	CompareTest{ "", "a", -1 },
+	CompareTest{ "abc", "abc", 0 },
+	CompareTest{ "ab", "abc", -1 },
+	CompareTest{ "abc", "ab", 1 },
+	CompareTest{ "x", "ab", 1 },
+	CompareTest{ "ab", "x", -1 },
+	CompareTest{ "x", "a", 1 },
+	CompareTest{ "b", "x", -1 },
+}
+
+func TestCompare(t *testing.T) {
+	for i := 0; i < len(comparetests); i++ {
+		tt := comparetests[i];
+		a := io.StringBytes(tt.a);
+		b := io.StringBytes(tt.b);
+		cmp := Compare(a, b);
+		eql := Equal(a, b);
+		if cmp != tt.cmp {
+			t.Errorf(`Compare(%q, %q) = %v`, tt.a, tt.b, cmp);
+		}
+		if eql != (tt.cmp==0) {
+			t.Errorf(`Equal(%q, %q) = %v`, tt.a, tt.b, eql);
+		}
+	}
+}
+
+
+type ExplodeTest struct {
+	s string;
+	a []string;
+}
+var explodetests = []ExplodeTest {
+	ExplodeTest{ abcd,	[]string{"a", "b", "c", "d"} },
+	ExplodeTest{ faces,	[]string{"☺", "☻", "☹" } },
+}
+func TestExplode(t *testing.T) {
+	for i := 0; i < len(explodetests); i++ {
+		tt := explodetests[i];
+		a := Explode(io.StringBytes(tt.s));
+		result := arrayOfString(a);
+		if !eq(result, tt.a) {
+			t.Errorf(`Explode("%s") = %v; want %v`, tt.s, result, tt.a);
+			continue;
+		}
+		s := Join(a, []byte{});
+		if string(s) != tt.s {
+			t.Errorf(`Join(Explode("%s"), "") = "%s"`, tt.s, s);
+		}
+	}
+}
+
+
+type SplitTest struct {
+	s string;
+	sep string;
+	a []string;
+}
+var splittests = []SplitTest {
+	SplitTest{ abcd,	"a",	[]string{"", "bcd"} },
+	SplitTest{ abcd,	"z",	[]string{"abcd"} },
+	SplitTest{ abcd,	"",	[]string{"a", "b", "c", "d"} },
+	SplitTest{ commas,	",",	[]string{"1", "2", "3", "4"} },
+	SplitTest{ dots,	"...",	[]string{"1", ".2", ".3", ".4"} },
+	SplitTest{ faces,	"☹",	[]string{"☺☻", ""} },
+	SplitTest{ faces,	"~",	[]string{faces} },
+	SplitTest{ faces,	"",	[]string{"☺", "☻", "☹"} },
+}
+func TestSplit(t *testing.T) {
+	for i := 0; i < len(splittests); i++ {
+		tt := splittests[i];
+		a := Split(io.StringBytes(tt.s), io.StringBytes(tt.sep));
+		result := arrayOfString(a);
+		if !eq(result, tt.a) {
+			t.Errorf(`Split("%s", "%s") = %v; want %v`, tt.s, tt.sep, result, tt.a);
+			continue;
+		}
+		s := Join(a, io.StringBytes(tt.sep));
+		if string(s) != tt.s {
+			t.Errorf(`Join(Split("%s", "%s"), "%s") = "%s"`, tt.s, tt.sep, tt.sep, s);
+		}
+	}
+}