ires 0.1.3 → 0.1.4

data/ext/vendor/golang.org/x/text/unicode/norm/iter_test.go

@@ -0,0 +1,98 @@
+// Copyright 2011 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 norm
+
+import (
+	"strings"
+	"testing"
+)
+
+func doIterNorm(f Form, s string) []byte {
+	acc := []byte{}
+	i := Iter{}
+	i.InitString(f, s)
+	for !i.Done() {
+		acc = append(acc, i.Next()...)
+	}
+	return acc
+}
+
+func TestIterNext(t *testing.T) {
+	runNormTests(t, "IterNext", func(f Form, out []byte, s string) []byte {
+		return doIterNorm(f, string(append(out, s...)))
+	})
+}
+
+type SegmentTest struct {
+	in  string
+	out []string
+}
+
+var segmentTests = []SegmentTest{
+	{"\u1E0A\u0323a", []string{"\x44\u0323\u0307", "a", ""}},
+	{rep('a', segSize), append(strings.Split(rep('a', segSize), ""), "")},
+	{rep('a', segSize+2), append(strings.Split(rep('a', segSize+2), ""), "")},
+	{rep('a', segSize) + "\u0300aa",
+		append(strings.Split(rep('a', segSize-1), ""), "a\u0300", "a", "a", "")},
+
+	// U+0f73 is NOT treated as a starter as it is a modifier
+	{"a" + grave(29) + "\u0f73", []string{"a" + grave(29), cgj + "\u0f73"}},
+	{"a\u0f73", []string{"a\u0f73"}},
+
+	// U+ff9e is treated as a non-starter.
+	// TODO: should we? Note that this will only affect iteration, as whether
+	// or not we do so does not affect the normalization output and will either
+	// way result in consistent iteration output.
+	{"a" + grave(30) + "\uff9e", []string{"a" + grave(30), cgj + "\uff9e"}},
+	{"a\uff9e", []string{"a\uff9e"}},
+}
+
+var segmentTestsK = []SegmentTest{
+	{"\u3332", []string{"\u30D5", "\u30A1", "\u30E9", "\u30C3", "\u30C8\u3099", ""}},
+	// last segment of multi-segment decomposition needs normalization
+	{"\u3332\u093C", []string{"\u30D5", "\u30A1", "\u30E9", "\u30C3", "\u30C8\u093C\u3099", ""}},
+	{"\u320E", []string{"\x28", "\uAC00", "\x29"}},
+
+	// last segment should be copied to start of buffer.
+	{"\ufdfa", []string{"\u0635", "\u0644", "\u0649", " ", "\u0627", "\u0644", "\u0644", "\u0647", " ", "\u0639", "\u0644", "\u064a", "\u0647", " ", "\u0648", "\u0633", "\u0644", "\u0645", ""}},
+	{"\ufdfa" + grave(30), []string{"\u0635", "\u0644", "\u0649", " ", "\u0627", "\u0644", "\u0644", "\u0647", " ", "\u0639", "\u0644", "\u064a", "\u0647", " ", "\u0648", "\u0633", "\u0644", "\u0645" + grave(30), ""}},
+	{"\uFDFA" + grave(64), []string{"\u0635", "\u0644", "\u0649", " ", "\u0627", "\u0644", "\u0644", "\u0647", " ", "\u0639", "\u0644", "\u064a", "\u0647", " ", "\u0648", "\u0633", "\u0644", "\u0645" + grave(30), cgj + grave(30), cgj + grave(4), ""}},
+
+	// Hangul and Jamo are grouped together.
+	{"\uAC00", []string{"\u1100\u1161", ""}},
+	{"\uAC01", []string{"\u1100\u1161\u11A8", ""}},
+	{"\u1100\u1161", []string{"\u1100\u1161", ""}},
+}
+
+// Note that, by design, segmentation is equal for composing and decomposing forms.
+func TestIterSegmentation(t *testing.T) {
+	segmentTest(t, "SegmentTestD", NFD, segmentTests)
+	segmentTest(t, "SegmentTestC", NFC, segmentTests)
+	segmentTest(t, "SegmentTestKD", NFKD, segmentTestsK)
+	segmentTest(t, "SegmentTestKC", NFKC, segmentTestsK)
+}
+
+func segmentTest(t *testing.T, name string, f Form, tests []SegmentTest) {
+	iter := Iter{}
+	for i, tt := range tests {
+		iter.InitString(f, tt.in)
+		for j, seg := range tt.out {
+			if seg == "" {
+				if !iter.Done() {
+					res := string(iter.Next())
+					t.Errorf(`%s:%d:%d: expected Done()==true, found segment %+q`, name, i, j, res)
+				}
+				continue
+			}
+			if iter.Done() {
+				t.Errorf("%s:%d:%d: Done()==true, want false", name, i, j)
+			}
+			seg = f.String(seg)
+			if res := string(iter.Next()); res != seg {
+				t.Errorf(`%s:%d:%d" segment was %+q (%d); want %+q (%d)`, name, i, j, pc(res), len(res), pc(seg), len(seg))
+			}
+		}
+	}
+}

data/ext/vendor/golang.org/x/text/unicode/norm/maketables.go

@@ -0,0 +1,976 @@
+// Copyright 2011 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.
+
+// +build ignore
+
+// Normalization table generator.
+// Data read from the web.
+// See forminfo.go for a description of the trie values associated with each rune.
+
+package main
+
+import (
+	"bytes"
+	"flag"
+	"fmt"
+	"io"
+	"log"
+	"sort"
+	"strconv"
+	"strings"
+
+	"golang.org/x/text/internal/gen"
+	"golang.org/x/text/internal/triegen"
+	"golang.org/x/text/internal/ucd"
+)
+
+func main() {
+	gen.Init()
+	loadUnicodeData()
+	compactCCC()
+	loadCompositionExclusions()
+	completeCharFields(FCanonical)
+	completeCharFields(FCompatibility)
+	computeNonStarterCounts()
+	verifyComputed()
+	printChars()
+	testDerived()
+	printTestdata()
+	makeTables()
+}
+
+var (
+	tablelist = flag.String("tables",
+		"all",
+		"comma-separated list of which tables to generate; "+
+			"can be 'decomp', 'recomp', 'info' and 'all'")
+	test = flag.Bool("test",
+		false,
+		"test existing tables against DerivedNormalizationProps and generate test data for regression testing")
+	verbose = flag.Bool("verbose",
+		false,
+		"write data to stdout as it is parsed")
+)
+
+const MaxChar = 0x10FFFF // anything above this shouldn't exist
+
+// Quick Check properties of runes allow us to quickly
+// determine whether a rune may occur in a normal form.
+// For a given normal form, a rune may be guaranteed to occur
+// verbatim (QC=Yes), may or may not combine with another
+// rune (QC=Maybe), or may not occur (QC=No).
+type QCResult int
+
+const (
+	QCUnknown QCResult = iota
+	QCYes
+	QCNo
+	QCMaybe
+)
+
+func (r QCResult) String() string {
+	switch r {
+	case QCYes:
+		return "Yes"
+	case QCNo:
+		return "No"
+	case QCMaybe:
+		return "Maybe"
+	}
+	return "***UNKNOWN***"
+}
+
+const (
+	FCanonical     = iota // NFC or NFD
+	FCompatibility        // NFKC or NFKD
+	FNumberOfFormTypes
+)
+
+const (
+	MComposed   = iota // NFC or NFKC
+	MDecomposed        // NFD or NFKD
+	MNumberOfModes
+)
+
+// This contains only the properties we're interested in.
+type Char struct {
+	name          string
+	codePoint     rune  // if zero, this index is not a valid code point.
+	ccc           uint8 // canonical combining class
+	origCCC       uint8
+	excludeInComp bool // from CompositionExclusions.txt
+	compatDecomp  bool // it has a compatibility expansion
+
+	nTrailingNonStarters uint8
+	nLeadingNonStarters  uint8 // must be equal to trailing if non-zero
+
+	forms [FNumberOfFormTypes]FormInfo // For FCanonical and FCompatibility
+
+	state State
+}
+
+var chars = make([]Char, MaxChar+1)
+var cccMap = make(map[uint8]uint8)
+
+func (c Char) String() string {
+	buf := new(bytes.Buffer)
+
+	fmt.Fprintf(buf, "%U [%s]:\n", c.codePoint, c.name)
+	fmt.Fprintf(buf, "  ccc: %v\n", c.ccc)
+	fmt.Fprintf(buf, "  excludeInComp: %v\n", c.excludeInComp)
+	fmt.Fprintf(buf, "  compatDecomp: %v\n", c.compatDecomp)
+	fmt.Fprintf(buf, "  state: %v\n", c.state)
+	fmt.Fprintf(buf, "  NFC:\n")
+	fmt.Fprint(buf, c.forms[FCanonical])
+	fmt.Fprintf(buf, "  NFKC:\n")
+	fmt.Fprint(buf, c.forms[FCompatibility])
+
+	return buf.String()
+}
+
+// In UnicodeData.txt, some ranges are marked like this:
+//	3400;<CJK Ideograph Extension A, First>;Lo;0;L;;;;;N;;;;;
+//	4DB5;<CJK Ideograph Extension A, Last>;Lo;0;L;;;;;N;;;;;
+// parseCharacter keeps a state variable indicating the weirdness.
+type State int
+
+const (
+	SNormal State = iota // known to be zero for the type
+	SFirst
+	SLast
+	SMissing
+)
+
+var lastChar = rune('\u0000')
+
+func (c Char) isValid() bool {
+	return c.codePoint != 0 && c.state != SMissing
+}
+
+type FormInfo struct {
+	quickCheck [MNumberOfModes]QCResult // index: MComposed or MDecomposed
+	verified   [MNumberOfModes]bool     // index: MComposed or MDecomposed
+
+	combinesForward  bool // May combine with rune on the right
+	combinesBackward bool // May combine with rune on the left
+	isOneWay         bool // Never appears in result
+	inDecomp         bool // Some decompositions result in this char.
+	decomp           Decomposition
+	expandedDecomp   Decomposition
+}
+
+func (f FormInfo) String() string {
+	buf := bytes.NewBuffer(make([]byte, 0))
+
+	fmt.Fprintf(buf, "    quickCheck[C]: %v\n", f.quickCheck[MComposed])
+	fmt.Fprintf(buf, "    quickCheck[D]: %v\n", f.quickCheck[MDecomposed])
+	fmt.Fprintf(buf, "    cmbForward: %v\n", f.combinesForward)
+	fmt.Fprintf(buf, "    cmbBackward: %v\n", f.combinesBackward)
+	fmt.Fprintf(buf, "    isOneWay: %v\n", f.isOneWay)
+	fmt.Fprintf(buf, "    inDecomp: %v\n", f.inDecomp)
+	fmt.Fprintf(buf, "    decomposition: %X\n", f.decomp)
+	fmt.Fprintf(buf, "    expandedDecomp: %X\n", f.expandedDecomp)
+
+	return buf.String()
+}
+
+type Decomposition []rune
+
+func parseDecomposition(s string, skipfirst bool) (a []rune, err error) {
+	decomp := strings.Split(s, " ")
+	if len(decomp) > 0 && skipfirst {
+		decomp = decomp[1:]
+	}
+	for _, d := range decomp {
+		point, err := strconv.ParseUint(d, 16, 64)
+		if err != nil {
+			return a, err
+		}
+		a = append(a, rune(point))
+	}
+	return a, nil
+}
+
+func loadUnicodeData() {
+	f := gen.OpenUCDFile("UnicodeData.txt")
+	defer f.Close()
+	p := ucd.New(f)
+	for p.Next() {
+		r := p.Rune(ucd.CodePoint)
+		char := &chars[r]
+
+		char.ccc = uint8(p.Uint(ucd.CanonicalCombiningClass))
+		decmap := p.String(ucd.DecompMapping)
+
+		exp, err := parseDecomposition(decmap, false)
+		isCompat := false
+		if err != nil {
+			if len(decmap) > 0 {
+				exp, err = parseDecomposition(decmap, true)
+				if err != nil {
+					log.Fatalf(`%U: bad decomp |%v|: "%s"`, r, decmap, err)
+				}
+				isCompat = true
+			}
+		}
+
+		char.name = p.String(ucd.Name)
+		char.codePoint = r
+		char.forms[FCompatibility].decomp = exp
+		if !isCompat {
+			char.forms[FCanonical].decomp = exp
+		} else {
+			char.compatDecomp = true
+		}
+		if len(decmap) > 0 {
+			char.forms[FCompatibility].decomp = exp
+		}
+	}
+	if err := p.Err(); err != nil {
+		log.Fatal(err)
+	}
+}
+
+// compactCCC converts the sparse set of CCC values to a continguous one,
+// reducing the number of bits needed from 8 to 6.
+func compactCCC() {
+	m := make(map[uint8]uint8)
+	for i := range chars {
+		c := &chars[i]
+		m[c.ccc] = 0
+	}
+	cccs := []int{}
+	for v, _ := range m {
+		cccs = append(cccs, int(v))
+	}
+	sort.Ints(cccs)
+	for i, c := range cccs {
+		cccMap[uint8(i)] = uint8(c)
+		m[uint8(c)] = uint8(i)
+	}
+	for i := range chars {
+		c := &chars[i]
+		c.origCCC = c.ccc
+		c.ccc = m[c.ccc]
+	}
+	if len(m) >= 1<<6 {
+		log.Fatalf("too many difference CCC values: %d >= 64", len(m))
+	}
+}
+
+// CompositionExclusions.txt has form:
+// 0958    # ...
+// See http://unicode.org/reports/tr44/ for full explanation
+func loadCompositionExclusions() {
+	f := gen.OpenUCDFile("CompositionExclusions.txt")
+	defer f.Close()
+	p := ucd.New(f)
+	for p.Next() {
+		c := &chars[p.Rune(0)]
+		if c.excludeInComp {
+			log.Fatalf("%U: Duplicate entry in exclusions.", c.codePoint)
+		}
+		c.excludeInComp = true
+	}
+	if e := p.Err(); e != nil {
+		log.Fatal(e)
+	}
+}
+
+// hasCompatDecomp returns true if any of the recursive
+// decompositions contains a compatibility expansion.
+// In this case, the character may not occur in NFK*.
+func hasCompatDecomp(r rune) bool {
+	c := &chars[r]
+	if c.compatDecomp {
+		return true
+	}
+	for _, d := range c.forms[FCompatibility].decomp {
+		if hasCompatDecomp(d) {
+			return true
+		}
+	}
+	return false
+}
+
+// Hangul related constants.
+const (
+	HangulBase = 0xAC00
+	HangulEnd  = 0xD7A4 // hangulBase + Jamo combinations (19 * 21 * 28)
+
+	JamoLBase = 0x1100
+	JamoLEnd  = 0x1113
+	JamoVBase = 0x1161
+	JamoVEnd  = 0x1176
+	JamoTBase = 0x11A8
+	JamoTEnd  = 0x11C3
+
+	JamoLVTCount = 19 * 21 * 28
+	JamoTCount   = 28
+)
+
+func isHangul(r rune) bool {
+	return HangulBase <= r && r < HangulEnd
+}
+
+func isHangulWithoutJamoT(r rune) bool {
+	if !isHangul(r) {
+		return false
+	}
+	r -= HangulBase
+	return r < JamoLVTCount && r%JamoTCount == 0
+}
+
+func ccc(r rune) uint8 {
+	return chars[r].ccc
+}
+
+// Insert a rune in a buffer, ordered by Canonical Combining Class.
+func insertOrdered(b Decomposition, r rune) Decomposition {
+	n := len(b)
+	b = append(b, 0)
+	cc := ccc(r)
+	if cc > 0 {
+		// Use bubble sort.
+		for ; n > 0; n-- {
+			if ccc(b[n-1]) <= cc {
+				break
+			}
+			b[n] = b[n-1]
+		}
+	}
+	b[n] = r
+	return b
+}
+
+// Recursively decompose.
+func decomposeRecursive(form int, r rune, d Decomposition) Decomposition {
+	dcomp := chars[r].forms[form].decomp
+	if len(dcomp) == 0 {
+		return insertOrdered(d, r)
+	}
+	for _, c := range dcomp {
+		d = decomposeRecursive(form, c, d)
+	}
+	return d
+}
+
+func completeCharFields(form int) {
+	// Phase 0: pre-expand decomposition.
+	for i := range chars {
+		f := &chars[i].forms[form]
+		if len(f.decomp) == 0 {
+			continue
+		}
+		exp := make(Decomposition, 0)
+		for _, c := range f.decomp {
+			exp = decomposeRecursive(form, c, exp)
+		}
+		f.expandedDecomp = exp
+	}
+
+	// Phase 1: composition exclusion, mark decomposition.
+	for i := range chars {
+		c := &chars[i]
+		f := &c.forms[form]
+
+		// Marks script-specific exclusions and version restricted.
+		f.isOneWay = c.excludeInComp
+
+		// Singletons
+		f.isOneWay = f.isOneWay || len(f.decomp) == 1
+
+		// Non-starter decompositions
+		if len(f.decomp) > 1 {
+			chk := c.ccc != 0 || chars[f.decomp[0]].ccc != 0
+			f.isOneWay = f.isOneWay || chk
+		}
+
+		// Runes that decompose into more than two runes.
+		f.isOneWay = f.isOneWay || len(f.decomp) > 2
+
+		if form == FCompatibility {
+			f.isOneWay = f.isOneWay || hasCompatDecomp(c.codePoint)
+		}
+
+		for _, r := range f.decomp {
+			chars[r].forms[form].inDecomp = true
+		}
+	}
+
+	// Phase 2: forward and backward combining.
+	for i := range chars {
+		c := &chars[i]
+		f := &c.forms[form]
+
+		if !f.isOneWay && len(f.decomp) == 2 {
+			f0 := &chars[f.decomp[0]].forms[form]
+			f1 := &chars[f.decomp[1]].forms[form]
+			if !f0.isOneWay {
+				f0.combinesForward = true
+			}
+			if !f1.isOneWay {
+				f1.combinesBackward = true
+			}
+		}
+		if isHangulWithoutJamoT(rune(i)) {
+			f.combinesForward = true
+		}
+	}
+
+	// Phase 3: quick check values.
+	for i := range chars {
+		c := &chars[i]
+		f := &c.forms[form]
+
+		switch {
+		case len(f.decomp) > 0:
+			f.quickCheck[MDecomposed] = QCNo
+		case isHangul(rune(i)):
+			f.quickCheck[MDecomposed] = QCNo
+		default:
+			f.quickCheck[MDecomposed] = QCYes
+		}
+		switch {
+		case f.isOneWay:
+			f.quickCheck[MComposed] = QCNo
+		case (i & 0xffff00) == JamoLBase:
+			f.quickCheck[MComposed] = QCYes
+			if JamoLBase <= i && i < JamoLEnd {
+				f.combinesForward = true
+			}
+			if JamoVBase <= i && i < JamoVEnd {
+				f.quickCheck[MComposed] = QCMaybe
+				f.combinesBackward = true
+				f.combinesForward = true
+			}
+			if JamoTBase <= i && i < JamoTEnd {
+				f.quickCheck[MComposed] = QCMaybe
+				f.combinesBackward = true
+			}
+		case !f.combinesBackward:
+			f.quickCheck[MComposed] = QCYes
+		default:
+			f.quickCheck[MComposed] = QCMaybe
+		}
+	}
+}
+
+func computeNonStarterCounts() {
+	// Phase 4: leading and trailing non-starter count
+	for i := range chars {
+		c := &chars[i]
+
+		runes := []rune{rune(i)}
+		// We always use FCompatibility so that the CGJ insertion points do not
+		// change for repeated normalizations with different forms.
+		if exp := c.forms[FCompatibility].expandedDecomp; len(exp) > 0 {
+			runes = exp
+		}
+		// We consider runes that combine backwards to be non-starters for the
+		// purpose of Stream-Safe Text Processing.
+		for _, r := range runes {
+			if cr := &chars[r]; cr.ccc == 0 && !cr.forms[FCompatibility].combinesBackward {
+				break
+			}
+			c.nLeadingNonStarters++
+		}
+		for i := len(runes) - 1; i >= 0; i-- {
+			if cr := &chars[runes[i]]; cr.ccc == 0 && !cr.forms[FCompatibility].combinesBackward {
+				break
+			}
+			c.nTrailingNonStarters++
+		}
+		if c.nTrailingNonStarters > 3 {
+			log.Fatalf("%U: Decomposition with more than 3 (%d) trailing modifiers (%U)", i, c.nTrailingNonStarters, runes)
+		}
+
+		if isHangul(rune(i)) {
+			c.nTrailingNonStarters = 2
+			if isHangulWithoutJamoT(rune(i)) {
+				c.nTrailingNonStarters = 1
+			}
+		}
+
+		if l, t := c.nLeadingNonStarters, c.nTrailingNonStarters; l > 0 && l != t {
+			log.Fatalf("%U: number of leading and trailing non-starters should be equal (%d vs %d)", i, l, t)
+		}
+		if t := c.nTrailingNonStarters; t > 3 {
+			log.Fatalf("%U: number of trailing non-starters is %d > 3", t)
+		}
+	}
+}
+
+func printBytes(w io.Writer, b []byte, name string) {
+	fmt.Fprintf(w, "// %s: %d bytes\n", name, len(b))
+	fmt.Fprintf(w, "var %s = [...]byte {", name)
+	for i, c := range b {
+		switch {
+		case i%64 == 0:
+			fmt.Fprintf(w, "\n// Bytes %x - %x\n", i, i+63)
+		case i%8 == 0:
+			fmt.Fprintf(w, "\n")
+		}
+		fmt.Fprintf(w, "0x%.2X, ", c)
+	}
+	fmt.Fprint(w, "\n}\n\n")
+}
+
+// See forminfo.go for format.
+func makeEntry(f *FormInfo, c *Char) uint16 {
+	e := uint16(0)
+	if r := c.codePoint; HangulBase <= r && r < HangulEnd {
+		e |= 0x40
+	}
+	if f.combinesForward {
+		e |= 0x20
+	}
+	if f.quickCheck[MDecomposed] == QCNo {
+		e |= 0x4
+	}
+	switch f.quickCheck[MComposed] {
+	case QCYes:
+	case QCNo:
+		e |= 0x10
+	case QCMaybe:
+		e |= 0x18
+	default:
+		log.Fatalf("Illegal quickcheck value %v.", f.quickCheck[MComposed])
+	}
+	e |= uint16(c.nTrailingNonStarters)
+	return e
+}
+
+// decompSet keeps track of unique decompositions, grouped by whether
+// the decomposition is followed by a trailing and/or leading CCC.
+type decompSet [7]map[string]bool
+
+const (
+	normalDecomp = iota
+	firstMulti
+	firstCCC
+	endMulti
+	firstLeadingCCC
+	firstCCCZeroExcept
+	firstStarterWithNLead
+	lastDecomp
+)
+
+var cname = []string{"firstMulti", "firstCCC", "endMulti", "firstLeadingCCC", "firstCCCZeroExcept", "firstStarterWithNLead", "lastDecomp"}
+
+func makeDecompSet() decompSet {
+	m := decompSet{}
+	for i := range m {
+		m[i] = make(map[string]bool)
+	}
+	return m
+}
+func (m *decompSet) insert(key int, s string) {
+	m[key][s] = true
+}
+
+func printCharInfoTables(w io.Writer) int {
+	mkstr := func(r rune, f *FormInfo) (int, string) {
+		d := f.expandedDecomp
+		s := string([]rune(d))
+		if max := 1 << 6; len(s) >= max {
+			const msg = "%U: too many bytes in decomposition: %d >= %d"
+			log.Fatalf(msg, r, len(s), max)
+		}
+		head := uint8(len(s))
+		if f.quickCheck[MComposed] != QCYes {
+			head |= 0x40
+		}
+		if f.combinesForward {
+			head |= 0x80
+		}
+		s = string([]byte{head}) + s
+
+		lccc := ccc(d[0])
+		tccc := ccc(d[len(d)-1])
+		cc := ccc(r)
+		if cc != 0 && lccc == 0 && tccc == 0 {
+			log.Fatalf("%U: trailing and leading ccc are 0 for non-zero ccc %d", r, cc)
+		}
+		if tccc < lccc && lccc != 0 {
+			const msg = "%U: lccc (%d) must be <= tcc (%d)"
+			log.Fatalf(msg, r, lccc, tccc)
+		}
+		index := normalDecomp
+		nTrail := chars[r].nTrailingNonStarters
+		nLead := chars[r].nLeadingNonStarters
+		if tccc > 0 || lccc > 0 || nTrail > 0 {
+			tccc <<= 2
+			tccc |= nTrail
+			s += string([]byte{tccc})
+			index = endMulti
+			for _, r := range d[1:] {
+				if ccc(r) == 0 {
+					index = firstCCC
+				}
+			}
+			if lccc > 0 || nLead > 0 {
+				s += string([]byte{lccc})
+				if index == firstCCC {
+					log.Fatalf("%U: multi-segment decomposition not supported for decompositions with leading CCC != 0", r)
+				}
+				index = firstLeadingCCC
+			}
+			if cc != lccc {
+				if cc != 0 {
+					log.Fatalf("%U: for lccc != ccc, expected ccc to be 0; was %d", r, cc)
+				}
+				index = firstCCCZeroExcept
+			}
+		} else if len(d) > 1 {
+			index = firstMulti
+		}
+		return index, s
+	}
+
+	decompSet := makeDecompSet()
+	const nLeadStr = "\x00\x01" // 0-byte length and tccc with nTrail.
+	decompSet.insert(firstStarterWithNLead, nLeadStr)
+
+	// Store the uniqued decompositions in a byte buffer,
+	// preceded by their byte length.
+	for _, c := range chars {
+		for _, f := range c.forms {
+			if len(f.expandedDecomp) == 0 {
+				continue
+			}
+			if f.combinesBackward {
+				log.Fatalf("%U: combinesBackward and decompose", c.codePoint)
+			}
+			index, s := mkstr(c.codePoint, &f)
+			decompSet.insert(index, s)
+		}
+	}
+
+	decompositions := bytes.NewBuffer(make([]byte, 0, 10000))
+	size := 0
+	positionMap := make(map[string]uint16)
+	decompositions.WriteString("\000")
+	fmt.Fprintln(w, "const (")
+	for i, m := range decompSet {
+		sa := []string{}
+		for s := range m {
+			sa = append(sa, s)
+		}
+		sort.Strings(sa)
+		for _, s := range sa {
+			p := decompositions.Len()
+			decompositions.WriteString(s)
+			positionMap[s] = uint16(p)
+		}
+		if cname[i] != "" {
+			fmt.Fprintf(w, "%s = 0x%X\n", cname[i], decompositions.Len())
+		}
+	}
+	fmt.Fprintln(w, "maxDecomp = 0x8000")
+	fmt.Fprintln(w, ")")
+	b := decompositions.Bytes()
+	printBytes(w, b, "decomps")
+	size += len(b)
+
+	varnames := []string{"nfc", "nfkc"}
+	for i := 0; i < FNumberOfFormTypes; i++ {
+		trie := triegen.NewTrie(varnames[i])
+
+		for r, c := range chars {
+			f := c.forms[i]
+			d := f.expandedDecomp
+			if len(d) != 0 {
+				_, key := mkstr(c.codePoint, &f)
+				trie.Insert(rune(r), uint64(positionMap[key]))
+				if c.ccc != ccc(d[0]) {
+					// We assume the lead ccc of a decomposition !=0 in this case.
+					if ccc(d[0]) == 0 {
+						log.Fatalf("Expected leading CCC to be non-zero; ccc is %d", c.ccc)
+					}
+				}
+			} else if c.nLeadingNonStarters > 0 && len(f.expandedDecomp) == 0 && c.ccc == 0 && !f.combinesBackward {
+				// Handle cases where it can't be detected that the nLead should be equal
+				// to nTrail.
+				trie.Insert(c.codePoint, uint64(positionMap[nLeadStr]))
+			} else if v := makeEntry(&f, &c)<<8 | uint16(c.ccc); v != 0 {
+				trie.Insert(c.codePoint, uint64(0x8000|v))
+			}
+		}
+		sz, err := trie.Gen(w, triegen.Compact(&normCompacter{name: varnames[i]}))
+		if err != nil {
+			log.Fatal(err)
+		}
+		size += sz
+	}
+	return size
+}
+
+func contains(sa []string, s string) bool {
+	for _, a := range sa {
+		if a == s {
+			return true
+		}
+	}
+	return false
+}
+
+func makeTables() {
+	w := &bytes.Buffer{}
+
+	size := 0
+	if *tablelist == "" {
+		return
+	}
+	list := strings.Split(*tablelist, ",")
+	if *tablelist == "all" {
+		list = []string{"recomp", "info"}
+	}
+
+	// Compute maximum decomposition size.
+	max := 0
+	for _, c := range chars {
+		if n := len(string(c.forms[FCompatibility].expandedDecomp)); n > max {
+			max = n
+		}
+	}
+
+	fmt.Fprintln(w, "const (")
+	fmt.Fprintln(w, "\t// Version is the Unicode edition from which the tables are derived.")
+	fmt.Fprintf(w, "\tVersion = %q\n", gen.UnicodeVersion())
+	fmt.Fprintln(w)
+	fmt.Fprintln(w, "\t// MaxTransformChunkSize indicates the maximum number of bytes that Transform")
+	fmt.Fprintln(w, "\t// may need to write atomically for any Form. Making a destination buffer at")
+	fmt.Fprintln(w, "\t// least this size ensures that Transform can always make progress and that")
+	fmt.Fprintln(w, "\t// the user does not need to grow the buffer on an ErrShortDst.")
+	fmt.Fprintf(w, "\tMaxTransformChunkSize = %d+maxNonStarters*4\n", len(string(0x034F))+max)
+	fmt.Fprintln(w, ")\n")
+
+	// Print the CCC remap table.
+	size += len(cccMap)
+	fmt.Fprintf(w, "var ccc = [%d]uint8{", len(cccMap))
+	for i := 0; i < len(cccMap); i++ {
+		if i%8 == 0 {
+			fmt.Fprintln(w)
+		}
+		fmt.Fprintf(w, "%3d, ", cccMap[uint8(i)])
+	}
+	fmt.Fprintln(w, "\n}\n")
+
+	if contains(list, "info") {
+		size += printCharInfoTables(w)
+	}
+
+	if contains(list, "recomp") {
+		// Note that we use 32 bit keys, instead of 64 bit.
+		// This clips the bits of three entries, but we know
+		// this won't cause a collision. The compiler will catch
+		// any changes made to UnicodeData.txt that introduces
+		// a collision.
+		// Note that the recomposition map for NFC and NFKC
+		// are identical.
+
+		// Recomposition map
+		nrentries := 0
+		for _, c := range chars {
+			f := c.forms[FCanonical]
+			if !f.isOneWay && len(f.decomp) > 0 {
+				nrentries++
+			}
+		}
+		sz := nrentries * 8
+		size += sz
+		fmt.Fprintf(w, "// recompMap: %d bytes (entries only)\n", sz)
+		fmt.Fprintln(w, "var recompMap = map[uint32]rune{")
+		for i, c := range chars {
+			f := c.forms[FCanonical]
+			d := f.decomp
+			if !f.isOneWay && len(d) > 0 {
+				key := uint32(uint16(d[0]))<<16 + uint32(uint16(d[1]))
+				fmt.Fprintf(w, "0x%.8X: 0x%.4X,\n", key, i)
+			}
+		}
+		fmt.Fprintf(w, "}\n\n")
+	}
+
+	fmt.Fprintf(w, "// Total size of tables: %dKB (%d bytes)\n", (size+512)/1024, size)
+	gen.WriteGoFile("tables.go", "norm", w.Bytes())
+}
+
+func printChars() {
+	if *verbose {
+		for _, c := range chars {
+			if !c.isValid() || c.state == SMissing {
+				continue
+			}
+			fmt.Println(c)
+		}
+	}
+}
+
+// verifyComputed does various consistency tests.
+func verifyComputed() {
+	for i, c := range chars {
+		for _, f := range c.forms {
+			isNo := (f.quickCheck[MDecomposed] == QCNo)
+			if (len(f.decomp) > 0) != isNo && !isHangul(rune(i)) {
+				log.Fatalf("%U: NF*D QC must be No if rune decomposes", i)
+			}
+
+			isMaybe := f.quickCheck[MComposed] == QCMaybe
+			if f.combinesBackward != isMaybe {
+				log.Fatalf("%U: NF*C QC must be Maybe if combinesBackward", i)
+			}
+			if len(f.decomp) > 0 && f.combinesForward && isMaybe {
+				log.Fatalf("%U: NF*C QC must be Yes or No if combinesForward and decomposes", i)
+			}
+
+			if len(f.expandedDecomp) != 0 {
+				continue
+			}
+			if a, b := c.nLeadingNonStarters > 0, (c.ccc > 0 || f.combinesBackward); a != b {
+				// We accept these runes to be treated differently (it only affects
+				// segment breaking in iteration, most likely on improper use), but
+				// reconsider if more characters are added.
+				// U+FF9E HALFWIDTH KATAKANA VOICED SOUND MARK;Lm;0;L;<narrow> 3099;;;;N;;;;;
+				// U+FF9F HALFWIDTH KATAKANA SEMI-VOICED SOUND MARK;Lm;0;L;<narrow> 309A;;;;N;;;;;
+				// U+3133 HANGUL LETTER KIYEOK-SIOS;Lo;0;L;<compat> 11AA;;;;N;HANGUL LETTER GIYEOG SIOS;;;;
+				// U+318E HANGUL LETTER ARAEAE;Lo;0;L;<compat> 11A1;;;;N;HANGUL LETTER ALAE AE;;;;
+				// U+FFA3 HALFWIDTH HANGUL LETTER KIYEOK-SIOS;Lo;0;L;<narrow> 3133;;;;N;HALFWIDTH HANGUL LETTER GIYEOG SIOS;;;;
+				// U+FFDC HALFWIDTH HANGUL LETTER I;Lo;0;L;<narrow> 3163;;;;N;;;;;
+				if i != 0xFF9E && i != 0xFF9F && !(0x3133 <= i && i <= 0x318E) && !(0xFFA3 <= i && i <= 0xFFDC) {
+					log.Fatalf("%U: nLead was %v; want %v", i, a, b)
+				}
+			}
+		}
+		nfc := c.forms[FCanonical]
+		nfkc := c.forms[FCompatibility]
+		if nfc.combinesBackward != nfkc.combinesBackward {
+			log.Fatalf("%U: Cannot combine combinesBackward\n", c.codePoint)
+		}
+	}
+}
+
+// Use values in DerivedNormalizationProps.txt to compare against the
+// values we computed.
+// DerivedNormalizationProps.txt has form:
+// 00C0..00C5    ; NFD_QC; N # ...
+// 0374          ; NFD_QC; N # ...
+// See http://unicode.org/reports/tr44/ for full explanation
+func testDerived() {
+	f := gen.OpenUCDFile("DerivedNormalizationProps.txt")
+	defer f.Close()
+	p := ucd.New(f)
+	for p.Next() {
+		r := p.Rune(0)
+		c := &chars[r]
+
+		var ftype, mode int
+		qt := p.String(1)
+		switch qt {
+		case "NFC_QC":
+			ftype, mode = FCanonical, MComposed
+		case "NFD_QC":
+			ftype, mode = FCanonical, MDecomposed
+		case "NFKC_QC":
+			ftype, mode = FCompatibility, MComposed
+		case "NFKD_QC":
+			ftype, mode = FCompatibility, MDecomposed
+		default:
+			continue
+		}
+		var qr QCResult
+		switch p.String(2) {
+		case "Y":
+			qr = QCYes
+		case "N":
+			qr = QCNo
+		case "M":
+			qr = QCMaybe
+		default:
+			log.Fatalf(`Unexpected quick check value "%s"`, p.String(2))
+		}
+		if got := c.forms[ftype].quickCheck[mode]; got != qr {
+			log.Printf("%U: FAILED %s (was %v need %v)\n", r, qt, got, qr)
+		}
+		c.forms[ftype].verified[mode] = true
+	}
+	if err := p.Err(); err != nil {
+		log.Fatal(err)
+	}
+	// Any unspecified value must be QCYes. Verify this.
+	for i, c := range chars {
+		for j, fd := range c.forms {
+			for k, qr := range fd.quickCheck {
+				if !fd.verified[k] && qr != QCYes {
+					m := "%U: FAIL F:%d M:%d (was %v need Yes) %s\n"
+					log.Printf(m, i, j, k, qr, c.name)
+				}
+			}
+		}
+	}
+}
+
+var testHeader = `const (
+	Yes = iota
+	No
+	Maybe
+)
+
+type formData struct {
+	qc              uint8
+	combinesForward bool
+	decomposition   string
+}
+
+type runeData struct {
+	r      rune
+	ccc    uint8
+	nLead  uint8
+	nTrail uint8
+	f      [2]formData // 0: canonical; 1: compatibility
+}
+
+func f(qc uint8, cf bool, dec string) [2]formData {
+	return [2]formData{{qc, cf, dec}, {qc, cf, dec}}
+}
+
+func g(qc, qck uint8, cf, cfk bool, d, dk string) [2]formData {
+	return [2]formData{{qc, cf, d}, {qck, cfk, dk}}
+}
+
+var testData = []runeData{
+`
+
+func printTestdata() {
+	type lastInfo struct {
+		ccc    uint8
+		nLead  uint8
+		nTrail uint8
+		f      string
+	}
+
+	last := lastInfo{}
+	w := &bytes.Buffer{}
+	fmt.Fprintf(w, testHeader)
+	for r, c := range chars {
+		f := c.forms[FCanonical]
+		qc, cf, d := f.quickCheck[MComposed], f.combinesForward, string(f.expandedDecomp)
+		f = c.forms[FCompatibility]
+		qck, cfk, dk := f.quickCheck[MComposed], f.combinesForward, string(f.expandedDecomp)
+		s := ""
+		if d == dk && qc == qck && cf == cfk {
+			s = fmt.Sprintf("f(%s, %v, %q)", qc, cf, d)
+		} else {
+			s = fmt.Sprintf("g(%s, %s, %v, %v, %q, %q)", qc, qck, cf, cfk, d, dk)
+		}
+		current := lastInfo{c.ccc, c.nLeadingNonStarters, c.nTrailingNonStarters, s}
+		if last != current {
+			fmt.Fprintf(w, "\t{0x%x, %d, %d, %d, %s},\n", r, c.origCCC, c.nLeadingNonStarters, c.nTrailingNonStarters, s)
+			last = current
+		}
+	}
+	fmt.Fprintln(w, "}")
+	gen.WriteGoFile("data_test.go", "norm", w.Bytes())
+}