summaryrefslogtreecommitdiffstats
path: root/vendor/golang.org/x/text
diff options
context:
space:
mode:
authorJesse Duffield <jessedduffield@gmail.com>2020-01-12 13:28:30 +1100
committerJesse Duffield <jessedduffield@gmail.com>2020-01-12 13:55:14 +1100
commit59ae1e159970e8adcfb4a228a93d621418b60483 (patch)
tree29cfc894ca4a8cc879fe70753a903e607c3b845a /vendor/golang.org/x/text
parent77a82e9d51aa789e51fc9ecd863187cc82aedf61 (diff)
bump gocui
Diffstat (limited to 'vendor/golang.org/x/text')
-rw-r--r--vendor/golang.org/x/text/internal/language/compact/gen.go64
-rw-r--r--vendor/golang.org/x/text/internal/language/compact/gen_index.go113
-rw-r--r--vendor/golang.org/x/text/internal/language/compact/gen_parents.go54
-rw-r--r--vendor/golang.org/x/text/internal/language/gen.go1520
-rw-r--r--vendor/golang.org/x/text/internal/language/gen_common.go20
-rw-r--r--vendor/golang.org/x/text/language/gen.go305
-rw-r--r--vendor/golang.org/x/text/unicode/norm/maketables.go986
-rw-r--r--vendor/golang.org/x/text/unicode/norm/triegen.go117
8 files changed, 3179 insertions, 0 deletions
diff --git a/vendor/golang.org/x/text/internal/language/compact/gen.go b/vendor/golang.org/x/text/internal/language/compact/gen.go
new file mode 100644
index 000000000..0c36a052f
--- /dev/null
+++ b/vendor/golang.org/x/text/internal/language/compact/gen.go
@@ -0,0 +1,64 @@
+// Copyright 2013 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
+
+// Language tag table generator.
+// Data read from the web.
+
+package main
+
+import (
+ "flag"
+ "fmt"
+ "log"
+
+ "golang.org/x/text/internal/gen"
+ "golang.org/x/text/unicode/cldr"
+)
+
+var (
+ test = flag.Bool("test",
+ false,
+ "test existing tables; can be used to compare web data with package data.")
+ outputFile = flag.String("output",
+ "tables.go",
+ "output file for generated tables")
+)
+
+func main() {
+ gen.Init()
+
+ w := gen.NewCodeWriter()
+ defer w.WriteGoFile("tables.go", "compact")
+
+ fmt.Fprintln(w, `import "golang.org/x/text/internal/language"`)
+
+ b := newBuilder(w)
+ gen.WriteCLDRVersion(w)
+
+ b.writeCompactIndex()
+}
+
+type builder struct {
+ w *gen.CodeWriter
+ data *cldr.CLDR
+ supp *cldr.SupplementalData
+}
+
+func newBuilder(w *gen.CodeWriter) *builder {
+ r := gen.OpenCLDRCoreZip()
+ defer r.Close()
+ d := &cldr.Decoder{}
+ data, err := d.DecodeZip(r)
+ if err != nil {
+ log.Fatal(err)
+ }
+ b := builder{
+ w: w,
+ data: data,
+ supp: data.Supplemental(),
+ }
+ return &b
+}
diff --git a/vendor/golang.org/x/text/internal/language/compact/gen_index.go b/vendor/golang.org/x/text/internal/language/compact/gen_index.go
new file mode 100644
index 000000000..136cefaf0
--- /dev/null
+++ b/vendor/golang.org/x/text/internal/language/compact/gen_index.go
@@ -0,0 +1,113 @@
+// Copyright 2015 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
+
+package main
+
+// This file generates derivative tables based on the language package itself.
+
+import (
+ "fmt"
+ "log"
+ "sort"
+ "strings"
+
+ "golang.org/x/text/internal/language"
+)
+
+// Compact indices:
+// Note -va-X variants only apply to localization variants.
+// BCP variants only ever apply to language.
+// The only ambiguity between tags is with regions.
+
+func (b *builder) writeCompactIndex() {
+ // Collect all language tags for which we have any data in CLDR.
+ m := map[language.Tag]bool{}
+ for _, lang := range b.data.Locales() {
+ // We include all locales unconditionally to be consistent with en_US.
+ // We want en_US, even though it has no data associated with it.
+
+ // TODO: put any of the languages for which no data exists at the end
+ // of the index. This allows all components based on ICU to use that
+ // as the cutoff point.
+ // if x := data.RawLDML(lang); false ||
+ // x.LocaleDisplayNames != nil ||
+ // x.Characters != nil ||
+ // x.Delimiters != nil ||
+ // x.Measurement != nil ||
+ // x.Dates != nil ||
+ // x.Numbers != nil ||
+ // x.Units != nil ||
+ // x.ListPatterns != nil ||
+ // x.Collations != nil ||
+ // x.Segmentations != nil ||
+ // x.Rbnf != nil ||
+ // x.Annotations != nil ||
+ // x.Metadata != nil {
+
+ // TODO: support POSIX natively, albeit non-standard.
+ tag := language.Make(strings.Replace(lang, "_POSIX", "-u-va-posix", 1))
+ m[tag] = true
+ // }
+ }
+
+ // TODO: plural rules are also defined for the deprecated tags:
+ // iw mo sh tl
+ // Consider removing these as compact tags.
+
+ // Include locales for plural rules, which uses a different structure.
+ for _, plurals := range b.supp.Plurals {
+ for _, rules := range plurals.PluralRules {
+ for _, lang := range strings.Split(rules.Locales, " ") {
+ m[language.Make(lang)] = true
+ }
+ }
+ }
+
+ var coreTags []language.CompactCoreInfo
+ var special []string
+
+ for t := range m {
+ if x := t.Extensions(); len(x) != 0 && fmt.Sprint(x) != "[u-va-posix]" {
+ log.Fatalf("Unexpected extension %v in %v", x, t)
+ }
+ if len(t.Variants()) == 0 && len(t.Extensions()) == 0 {
+ cci, ok := language.GetCompactCore(t)
+ if !ok {
+ log.Fatalf("Locale for non-basic language %q", t)
+ }
+ coreTags = append(coreTags, cci)
+ } else {
+ special = append(special, t.String())
+ }
+ }
+
+ w := b.w
+
+ sort.Slice(coreTags, func(i, j int) bool { return coreTags[i] < coreTags[j] })
+ sort.Strings(special)
+
+ w.WriteComment(`
+ NumCompactTags is the number of common tags. The maximum tag is
+ NumCompactTags-1.`)
+ w.WriteConst("NumCompactTags", len(m))
+
+ fmt.Fprintln(w, "const (")
+ for i, t := range coreTags {
+ fmt.Fprintf(w, "%s ID = %d\n", ident(t.Tag().String()), i)
+ }
+ for i, t := range special {
+ fmt.Fprintf(w, "%s ID = %d\n", ident(t), i+len(coreTags))
+ }
+ fmt.Fprintln(w, ")")
+
+ w.WriteVar("coreTags", coreTags)
+
+ w.WriteConst("specialTagsStr", strings.Join(special, " "))
+}
+
+func ident(s string) string {
+ return strings.Replace(s, "-", "", -1) + "Index"
+}
diff --git a/vendor/golang.org/x/text/internal/language/compact/gen_parents.go b/vendor/golang.org/x/text/internal/language/compact/gen_parents.go
new file mode 100644
index 000000000..9543d5832
--- /dev/null
+++ b/vendor/golang.org/x/text/internal/language/compact/gen_parents.go
@@ -0,0 +1,54 @@
+// Copyright 2018 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
+
+package main
+
+import (
+ "log"
+
+ "golang.org/x/text/internal/gen"
+ "golang.org/x/text/internal/language"
+ "golang.org/x/text/internal/language/compact"
+ "golang.org/x/text/unicode/cldr"
+)
+
+func main() {
+ r := gen.OpenCLDRCoreZip()
+ defer r.Close()
+
+ d := &cldr.Decoder{}
+ data, err := d.DecodeZip(r)
+ if err != nil {
+ log.Fatalf("DecodeZip: %v", err)
+ }
+
+ w := gen.NewCodeWriter()
+ defer w.WriteGoFile("parents.go", "compact")
+
+ // Create parents table.
+ type ID uint16
+ parents := make([]ID, compact.NumCompactTags)
+ for _, loc := range data.Locales() {
+ tag := language.MustParse(loc)
+ index, ok := compact.FromTag(tag)
+ if !ok {
+ continue
+ }
+ parentIndex := compact.ID(0) // und
+ for p := tag.Parent(); p != language.Und; p = p.Parent() {
+ if x, ok := compact.FromTag(p); ok {
+ parentIndex = x
+ break
+ }
+ }
+ parents[index] = ID(parentIndex)
+ }
+
+ w.WriteComment(`
+ parents maps a compact index of a tag to the compact index of the parent of
+ this tag.`)
+ w.WriteVar("parents", parents)
+}
diff --git a/vendor/golang.org/x/text/internal/language/gen.go b/vendor/golang.org/x/text/internal/language/gen.go
new file mode 100644
index 000000000..cdcc7febc
--- /dev/null
+++ b/vendor/golang.org/x/text/internal/language/gen.go
@@ -0,0 +1,1520 @@
+// Copyright 2013 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
+
+// Language tag table generator.
+// Data read from the web.
+
+package main
+
+import (
+ "bufio"
+ "flag"
+ "fmt"
+ "io"
+ "io/ioutil"
+ "log"
+ "math"
+ "reflect"
+ "regexp"
+ "sort"
+ "strconv"
+ "strings"
+
+ "golang.org/x/text/internal/gen"
+ "golang.org/x/text/internal/tag"
+ "golang.org/x/text/unicode/cldr"
+)
+
+var (
+ test = flag.Bool("test",
+ false,
+ "test existing tables; can be used to compare web data with package data.")
+ outputFile = flag.String("output",
+ "tables.go",
+ "output file for generated tables")
+)
+
+var comment = []string{
+ `
+lang holds an alphabetically sorted list of ISO-639 language identifiers.
+All entries are 4 bytes. The index of the identifier (divided by 4) is the language tag.
+For 2-byte language identifiers, the two successive bytes have the following meaning:
+ - if the first letter of the 2- and 3-letter ISO codes are the same:
+ the second and third letter of the 3-letter ISO code.
+ - otherwise: a 0 and a by 2 bits right-shifted index into altLangISO3.
+For 3-byte language identifiers the 4th byte is 0.`,
+ `
+langNoIndex is a bit vector of all 3-letter language codes that are not used as an index
+in lookup tables. The language ids for these language codes are derived directly
+from the letters and are not consecutive.`,
+ `
+altLangISO3 holds an alphabetically sorted list of 3-letter language code alternatives
+to 2-letter language codes that cannot be derived using the method described above.
+Each 3-letter code is followed by its 1-byte langID.`,
+ `
+altLangIndex is used to convert indexes in altLangISO3 to langIDs.`,
+ `
+AliasMap maps langIDs to their suggested replacements.`,
+ `
+script is an alphabetically sorted list of ISO 15924 codes. The index
+of the script in the string, divided by 4, is the internal scriptID.`,
+ `
+isoRegionOffset needs to be added to the index of regionISO to obtain the regionID
+for 2-letter ISO codes. (The first isoRegionOffset regionIDs are reserved for
+the UN.M49 codes used for groups.)`,
+ `
+regionISO holds a list of alphabetically sorted 2-letter ISO region codes.
+Each 2-letter codes is followed by two bytes with the following meaning:
+ - [A-Z}{2}: the first letter of the 2-letter code plus these two
+ letters form the 3-letter ISO code.
+ - 0, n: index into altRegionISO3.`,
+ `
+regionTypes defines the status of a region for various standards.`,
+ `
+m49 maps regionIDs to UN.M49 codes. The first isoRegionOffset entries are
+codes indicating collections of regions.`,
+ `
+m49Index gives indexes into fromM49 based on the three most significant bits
+of a 10-bit UN.M49 code. To search an UN.M49 code in fromM49, search in
+ fromM49[m49Index[msb39(code)]:m49Index[msb3(code)+1]]
+for an entry where the first 7 bits match the 7 lsb of the UN.M49 code.
+The region code is stored in the 9 lsb of the indexed value.`,
+ `
+fromM49 contains entries to map UN.M49 codes to regions. See m49Index for details.`,
+ `
+altRegionISO3 holds a list of 3-letter region codes that cannot be
+mapped to 2-letter codes using the default algorithm. This is a short list.`,
+ `
+altRegionIDs holds a list of regionIDs the positions of which match those
+of the 3-letter ISO codes in altRegionISO3.`,
+ `
+variantNumSpecialized is the number of specialized variants in variants.`,
+ `
+suppressScript is an index from langID to the dominant script for that language,
+if it exists. If a script is given, it should be suppressed from the language tag.`,
+ `
+likelyLang is a lookup table, indexed by langID, for the most likely
+scripts and regions given incomplete information. If more entries exist for a
+given language, region and script are the index and size respectively
+of the list in likelyLangList.`,
+ `
+likelyLangList holds lists info associated with likelyLang.`,
+ `
+likelyRegion is a lookup table, indexed by regionID, for the most likely
+languages and scripts given incomplete information. If more entries exist
+for a given regionID, lang and script are the index and size respectively
+of the list in likelyRegionList.
+TODO: exclude containers and user-definable regions from the list.`,
+ `
+likelyRegionList holds lists info associated with likelyRegion.`,
+ `
+likelyScript is a lookup table, indexed by scriptID, for the most likely
+languages and regions given a script.`,
+ `
+nRegionGroups is the number of region groups.`,
+ `
+regionInclusion maps region identifiers to sets of regions in regionInclusionBits,
+where each set holds all groupings that are directly connected in a region
+containment graph.`,
+ `
+regionInclusionBits is an array of bit vectors where every vector represents
+a set of region groupings. These sets are used to compute the distance
+between two regions for the purpose of language matching.`,
+ `
+regionInclusionNext marks, for each entry in regionInclusionBits, the set of
+all groups that are reachable from the groups set in the respective entry.`,
+}
+
+// TODO: consider changing some of these structures to tries. This can reduce
+// memory, but may increase the need for memory allocations. This could be
+// mitigated if we can piggyback on language tags for common cases.
+
+func failOnError(e error) {
+ if e != nil {
+ log.Panic(e)
+ }
+}
+
+type setType int
+
+const (
+ Indexed setType = 1 + iota // all elements must be of same size
+ Linear
+)
+
+type stringSet struct {
+ s []string
+ sorted, frozen bool
+
+ // We often need to update values after the creation of an index is completed.
+ // We include a convenience map for keeping track of this.
+ update map[string]string
+ typ setType // used for checking.
+}
+
+func (ss *stringSet) clone() stringSet {
+ c := *ss
+ c.s = append([]string(nil), c.s...)
+ return c
+}
+
+func (ss *stringSet) setType(t setType) {
+ if ss.typ != t && ss.typ != 0 {
+ log.Panicf("type %d cannot be assigned as it was already %d", t, ss.typ)
+ }
+}
+
+// parse parses a whitespace-separated string and initializes ss with its
+// components.
+func (ss *stringSet) parse(s string) {
+ scan := bufio.NewScanner(strings.NewReader(s))
+ scan.Split(bufio.ScanWords)
+ for scan.Scan() {
+ ss.add(scan.Text())
+ }
+}
+
+func (ss *stringSet) assertChangeable() {
+ if ss.frozen {
+ log.Panic("attempt to modify a frozen stringSet")
+ }
+}
+
+func (ss *stringSet) add(s string) {
+ ss.assertChangeable()
+ ss.s = append(ss.s, s)
+ ss.sorted = ss.frozen
+}
+
+func (ss *stringSet) freeze() {
+ ss.compact()
+ ss.frozen = true
+}
+
+func (ss *stringSet) compact() {
+ if ss.sorted {
+ return
+ }
+ a := ss.s
+ sort.Strings(a)
+ k := 0
+ for i := 1; i < len(a); i++ {
+ if a[k] != a[i] {
+ a[k+1] = a[i]
+ k++
+ }
+ }
+ ss.s = a[:k+1]
+ ss.sorted = ss.frozen
+}
+
+type funcSorter struct {
+ fn func(a, b string) bool
+ sort.StringSlice
+}
+
+func (s funcSorter) Less(i, j int) bool {
+ return s.fn(s.StringSlice[i], s.StringSlice[j])
+}
+
+func (ss *stringSet) sortFunc(f func(a, b string) bool) {
+ ss.compact()
+ sort.Sort(funcSorter{f, sort.StringSlice(ss.s)})
+}
+
+func (ss *stringSet) remove(s string) {
+ ss.assertChangeable()
+ if i, ok := ss.find(s); ok {
+ copy(ss.s[i:], ss.s[i+1:])
+ ss.s = ss.s[:len(ss.s)-1]
+ }
+}
+
+func (ss *stringSet) replace(ol, nu string) {
+ ss.s[ss.index(ol)] = nu
+ ss.sorted = ss.frozen
+}
+
+func (ss *stringSet) index(s string) int {
+ ss.setType(Indexed)
+ i, ok := ss.find(s)
+ if !ok {
+ if i < len(ss.s) {
+ log.Panicf("find: item %q is not in list. Closest match is %q.", s, ss.s[i])
+ }
+ log.Panicf("find: item %q is not in list", s)
+
+ }
+ return i
+}
+
+func (ss *stringSet) find(s string) (int, bool) {
+ ss.compact()
+ i := sort.SearchStrings(ss.s, s)
+ return i, i != len(ss.s) && ss.s[i] == s
+}
+
+func (ss *stringSet) slice() []string {
+ ss.compact()
+ return ss.s
+}
+
+func (ss *stringSet) updateLater(v, key string) {
+ if ss.update == nil {
+ ss.update = map[string]string{}
+ }
+ ss.update[v] = key
+}
+
+// join joins the string and ensures that all entries are of the same length.
+func (ss *stringSet) join() string {
+ ss.setType(Indexed)
+ n := len(ss.s[0])
+ for _, s := range ss.s {
+ if len(s) != n {
+ log.Panicf("join: not all entries are of the same length: %q", s)
+ }
+ }
+ ss.s = append(ss.s, strings.Repeat("\xff", n))
+ return strings.Join(ss.s, "")
+}
+
+// ianaEntry holds information for an entry in the IANA Language Subtag Repository.
+// All types use the same entry.
+// See http://tools.ietf.org/html/bcp47#section-5.1 for a description of the various
+// fields.
+type ianaEntry struct {
+ typ string
+ description []string
+ scope string
+ added string
+ preferred string
+ deprecated string
+ suppressScript string
+ macro string
+ prefix []string
+}
+
+type builder struct {
+ w *gen.CodeWriter
+ hw io.Writer // MultiWriter for w and w.Hash
+ data *cldr.CLDR
+ supp *cldr.SupplementalData
+
+ // indices
+ locale stringSet // common locales
+ lang stringSet // canonical language ids (2 or 3 letter ISO codes) with data
+ langNoIndex stringSet // 3-letter ISO codes with no associated data
+ script stringSet // 4-letter ISO codes
+ region stringSet // 2-letter ISO or 3-digit UN M49 codes
+ variant stringSet // 4-8-alphanumeric variant code.
+
+ // Region codes that are groups with their corresponding group IDs.
+ groups map[int]index
+
+ // langInfo
+ registry map[string]*ianaEntry
+}
+
+type index uint
+
+func newBuilder(w *gen.CodeWriter) *builder {
+ r := gen.OpenCLDRCoreZip()
+ defer r.Close()
+ d := &cldr.Decoder{}
+ data, err := d.DecodeZip(r)
+ failOnError(err)
+ b := builder{
+ w: w,
+ hw: io.MultiWriter(w, w.Hash),
+ data: data,
+ supp: data.Supplemental(),
+ }
+ b.parseRegistry()
+ return &b
+}
+
+func (b *builder) parseRegistry() {
+ r := gen.OpenIANAFile("assignments/language-subtag-registry")
+ defer r.Close()
+ b.registry = make(map[string]*ianaEntry)
+
+ scan := bufio.NewScanner(r)
+ scan.Split(bufio.ScanWords)
+ var record *ianaEntry
+ for more := scan.Scan(); more; {
+ key := scan.Text()
+ more = scan.Scan()
+ value := scan.Text()
+ switch key {
+ case "Type:":
+ record = &ianaEntry{typ: value}
+ case "Subtag:", "Tag:":
+ if s := strings.SplitN(value, "..", 2); len(s) > 1 {
+ for a := s[0]; a <= s[1]; a = inc(a) {
+ b.addToRegistry(a, record)
+ }
+ } else {
+ b.addToRegistry(value, record)
+ }
+ case "Suppress-Script:":
+ record.suppressScript = value
+ case "Added:":
+ record.added = value
+ case "Deprecated:":
+ record.deprecated = value
+ case "Macrolanguage:":
+ record.macro = value
+ case "Preferred-Value:":
+ record.preferred = value
+ case "Prefix:":
+ record.prefix = append(record.prefix, value)
+ case "Scope:":
+ record.scope = value
+ case "Description:":
+ buf := []byte(value)
+ for more = scan.Scan(); more; more = scan.Scan() {
+ b := scan.Bytes()
+ if b[0] == '%' || b[len(b)-1] == ':' {
+ break
+ }
+ buf = append(buf, ' ')
+ buf = append(buf, b...)
+ }
+ record.description = append(record.description, string(buf))
+ continue
+ default:
+ continue
+ }
+ more = scan.Scan()
+ }
+ if scan.Err() != nil {
+ log.Panic(scan.Err())
+ }
+}
+
+func (b *builder) addToRegistry(key string, entry *ianaEntry) {
+ if info, ok := b.registry[key]; ok {
+ if info.typ != "language" || entry.typ != "extlang" {
+ log.Fatalf("parseRegistry: tag %q already exists", key)
+ }
+ } else {
+ b.registry[key] = entry
+ }
+}
+
+var commentIndex = make(map[string]string)
+
+func init() {
+ for _, s := range comment {
+ key := strings.TrimSpace(strings.SplitN(s, " ", 2)[0])
+ commentIndex[key] = s
+ }
+}
+
+func (b *builder) comment(name string) {
+ if s := commentIndex[name]; len(s) > 0 {
+ b.w.WriteComment(s)
+ } else {
+ fmt.Fprintln(b.w)
+ }
+}
+
+func (b *builder) pf(f string, x ...interface{}) {
+ fmt.Fprintf(b.hw, f, x...)
+ fmt.Fprint(b.hw, "\n")
+}
+
+func (b *builder) p(x ...interface{}) {
+ fmt.Fprintln(b.hw, x...)
+}
+
+func (b *builder) addSize(s int) {
+ b.w.Size += s
+ b.pf("// Size: %d bytes", s)
+}
+
+func (b *builder) writeConst(name string, x interface{}) {
+ b.comment(name)
+ b.w.WriteConst(name, x)
+}
+
+// writeConsts computes f(v) for all v in values and writes the results
+// as constants named _v to a single constant block.
+func (b *builder) writeConsts(f func(string) int, values ...string) {
+ b.pf("const (")
+ for _, v := range values {
+ b.pf("\t_%s = %v", v, f(v))
+ }
+ b.pf(")")
+}
+
+// writeType writes the type of the given value, which must be a struct.
+func (b *builder) writeType(value interface{}) {
+ b.comment(reflect.TypeOf(value).Name())
+ b.w.WriteType(value)
+}
+
+func (b *builder) writeSlice(name string, ss interface{}) {
+ b.writeSliceAddSize(name, 0, ss)
+}
+
+func (b *builder) writeSliceAddSize(name string, extraSize int, ss interface{}) {
+ b.comment(name)
+ b.w.Size += extraSize
+ v := reflect.ValueOf(ss)
+ t := v.Type().Elem()
+ b.pf("// Size: %d bytes, %d elements", v.Len()*int(t.Size())+extraSize, v.Len())
+
+ fmt.Fprintf(b.w, "var %s = ", name)
+ b.w.WriteArray(ss)
+ b.p()
+}
+
+type FromTo struct {
+ From, To uint16
+}
+
+func (b *builder) writeSortedMap(name string, ss *stringSet, index func(s string) uint16) {
+ ss.sortFunc(func(a, b string) bool {
+ return index(a) < index(b)
+ })
+ m := []FromTo{}
+ for _, s := range ss.s {
+ m = append(m, FromTo{index(s), index(ss.update[s])})
+ }
+ b.writeSlice(name, m)
+}
+
+const base = 'z' - 'a' + 1
+
+func strToInt(s string) uint {
+ v := uint(0)
+ for i := 0; i < len(s); i++ {
+ v *= base
+ v += uint(s[i] - 'a')
+ }
+ return v
+}
+
+// converts the given integer to the original ASCII string passed to strToInt.
+// len(s) must match the number of characters obtained.
+func intToStr(v uint, s []byte) {
+ for i := len(s) - 1; i >= 0; i-- {
+ s[i] = byte(v%base) + 'a'
+ v /= base
+ }
+}
+
+func (b *builder) writeBitVector(name string, ss []string) {
+ vec := make([]uint8, int(math.Ceil(math.Pow(base, float64(len(ss[0])))/8)))
+ for _, s := range ss {
+ v := strToInt(s)
+ vec[v/8] |= 1 << (v % 8)
+ }
+ b.writeSlice(name, vec)
+}
+
+// TODO: convert this type into a list or two-stage trie.
+func (b *builder) writeMapFunc(name string, m map[string]string, f func(string) uint16) {
+ b.comment(name)
+ v := reflect.ValueOf(m)
+ sz := v.Len() * (2 + int(v.Type().Key().Size()))
+ for _, k := range m {
+ sz += len(k)
+ }
+ b.addSize(sz)
+ keys := []string{}
+ b.pf(`var %s = map[string]uint16{`, name)
+ for k := range m {
+ keys = append(keys, k)
+ }
+ sort.Strings(keys)
+ for _, k := range keys {
+ b.pf("\t%q: %v,", k, f(m[k]))
+ }
+ b.p("}")
+}
+
+func (b *builder) writeMap(name string, m interface{}) {
+ b.comment(name)
+ v := reflect.ValueOf(m)
+ sz := v.Len() * (2 + int(v.Type().Key().Size()) + int(v.Type().Elem().Size()))
+ b.addSize(sz)
+ f := strings.FieldsFunc(fmt.Sprintf("%#v", m), func(r rune) bool {
+ return strings.IndexRune("{}, ", r) != -1
+ })
+ sort.Strings(f[1:])
+ b.pf(`var %s = %s{`, name, f[0])
+ for _, kv := range f[1:] {
+ b.pf("\t%s,", kv)
+ }
+ b.p("}")
+}
+
+func (b *builder) langIndex(s string) uint16 {
+ if s == "und" {
+ return 0
+ }
+ if i, ok := b.lang.find(s); ok {
+ return uint16(i)
+ }
+ return uint16(strToInt(s)) + uint16(len(b.lang.s))
+}
+
+// inc advances the string to its lexicographical successor.
+func inc(s string) string {
+ const maxTagLength = 4
+ var buf [maxTagLength]byte
+ intToStr(strToInt(strings.ToLower(s))+1, buf[:len(s)])
+ for i := 0; i < len(s); i++ {
+ if s[i] <= 'Z' {
+ buf[i] -= 'a' - 'A'
+ }
+ }
+ return string(buf[:len(s)])
+}
+
+func (b *builder) parseIndices() {
+ meta := b.supp.Metadata
+
+ for k, v := range b.registry {
+ var ss *stringSet
+ switch v.typ {
+ case "language":
+ if len(k) == 2 || v.suppressScript != "" || v.scope == "special" {
+ b.lang.add(k)
+ continue
+ } else {
+ ss = &b.langNoIndex
+ }
+ case "region":
+ ss = &b.region
+ case "script":
+ ss = &b.script
+ case "variant":
+ ss = &b.variant
+ default:
+ continue
+ }
+ ss.add(k)
+ }
+ // Include any language for which there is data.
+ for _, lang := range b.data.Locales() {
+ if x := b.data.RawLDML(lang); false ||
+ x.LocaleDisplayNames != nil ||
+ x.Characters != nil ||
+ x.Delimiters != nil ||
+ x.Measurement != nil ||
+ x.Dates != nil ||
+ x.Numbers != nil ||
+ x.Units != nil ||
+ x.ListPatterns != nil ||
+ x.Collations != nil ||
+ x.Segmentations != nil ||
+ x.Rbnf != nil ||
+ x.Annotations != nil ||
+ x.Metadata != nil {
+
+ from := strings.Split(lang, "_")
+ if lang := from[0]; lang != "root" {
+ b.lang.add(lang)
+ }
+ }
+ }
+ // Include locales for plural rules, which uses a different structure.
+ for _, plurals := range b.data.Supplemental().Plurals {
+ for _, rules := range plurals.PluralRules {
+ for _, lang := range strings.Split(rules.Locales, " ") {
+ if lang = strings.Split(lang, "_")[0]; lang != "root" {
+ b.lang.add(lang)
+ }
+ }
+ }
+ }
+ // Include languages in likely subtags.
+ for _, m := range b.supp.LikelySubtags.LikelySubtag {
+ from := strings.Split(m.From, "_")
+ b.lang.add(from[0])
+ }
+ // Include ISO-639 alpha-3 bibliographic entries.
+ for _, a := range meta.Alias.LanguageAlias {
+ if a.Reason == "bibliographic" {
+ b.langNoIndex.add(a.Type)
+ }
+ }
+ // Include regions in territoryAlias (not all are in the IANA registry!)
+ for _, reg := range b.supp.Metadata.Alias.TerritoryAlias {
+ if len(reg.Type) == 2 {
+ b.region.add(reg.Type)
+ }
+ }
+
+ for _, s := range b.lang.s {
+ if len(s) == 3 {
+ b.langNoIndex.remove(s)
+ }
+ }
+ b.writeConst("NumLanguages", len(b.lang.slice())+len(b.langNoIndex.slice()))
+ b.writeConst("NumScripts", len(b.script.slice()))
+ b.writeConst("NumRegions", len(b.region.slice()))
+
+ // Add dummy codes at the start of each list to represent "unspecified".
+ b.lang.add("---")
+ b.script.add("----")
+ b.region.add("---")
+
+ // common locales
+ b.locale.parse(meta.DefaultContent.Locales)
+}
+
+// TODO: region inclusion data will probably not be use used in future matchers.
+
+func (b *builder) computeRegionGroups() {
+ b.groups = make(map[int]index)
+
+ // Create group indices.
+ for i := 1; b.region.s[i][0] < 'A'; i++ { // Base M49 indices on regionID.
+ b.groups[i] = index(len(b.groups))
+ }
+ for _, g := range b.supp.TerritoryContainment.Group {
+ // Skip UN and EURO zone as they are flattening the containment
+ // relationship.
+ if g.Type == "EZ" || g.Type == "UN" {
+ continue
+ }
+ group := b.region.index(g.Type)
+ if _, ok := b.groups[group]; !ok {
+ b.groups[group] = index(len(b.groups))
+ }
+ }
+ if len(b.groups) > 64 {
+ log.Fatalf("only 64 groups supported, found %d", len(b.groups))
+ }
+ b.writeConst("nRegionGroups", len(b.groups))
+}
+
+var langConsts = []string{
+ "af", "am", "ar", "az", "bg", "bn", "ca", "cs", "da", "de", "el", "en", "es",
+ "et", "fa", "fi", "fil", "fr", "gu", "he", "hi", "hr", "hu", "hy", "id", "is",
+ "it", "ja", "ka", "kk", "km", "kn", "ko", "ky", "lo", "lt", "lv", "mk", "ml",
+ "mn", "mo", "mr", "ms", "mul", "my", "nb", "ne", "nl", "no", "pa", "pl", "pt",
+ "ro", "ru", "sh", "si", "sk", "sl", "sq", "sr", "sv", "sw", "ta", "te", "th",
+ "tl", "tn", "tr", "uk", "ur", "uz", "vi", "zh", "zu",
+
+ // constants for grandfathered tags (if not already defined)
+ "jbo", "ami", "bnn", "hak", "tlh", "lb", "nv", "pwn", "tao", "tay", "tsu",
+ "nn", "sfb", "vgt", "sgg", "cmn", "nan", "hsn",
+}
+
+// writeLanguage generates all tables needed for language canonicalization.
+func (b *builder) writeLanguage() {
+ meta := b.supp.Metadata
+
+ b.writeConst("nonCanonicalUnd", b.lang.index("und"))
+ b.writeConsts(func(s string) int { return int(b.langIndex(s)) }, langConsts...)
+ b.writeConst("langPrivateStart", b.langIndex("qaa"))
+ b.writeConst("langPrivateEnd", b.langIndex("qtz"))
+
+ // Get language codes that need to be mapped (overlong 3-letter codes,
+ // deprecated 2-letter codes, legacy and grandfathered tags.)
+ langAliasMap := stringSet{}
+ aliasTypeMap := map[string]AliasType{}
+
+ // altLangISO3 get the alternative ISO3 names that need to be mapped.
+ altLangISO3 := stringSet{}
+ // Add dummy start to avoid the use of index 0.
+ altLangISO3.add("---")
+ altLangISO3.updateLater("---", "aa")
+
+ lang := b.lang.clone()
+ for _, a := range