homelab/traefik
1
0
Fork 0
Code Activity

Custom resource definition

Browse source 
Co-authored-by: Mathieu Lonjaret <mathieu.lonjaret@gmail.com>
This commit is contained in:
Ludovic Fernandez 2019-03-14 15:56:06 +01:00 • committed by Traefiker Bot
parent cfaf47c8a2
commit 4c060a78cc
1348 changed files with 92364 additions and 55766 deletions
Download patch file Download diff file Expand all files Collapse all files

162
vendor/golang.org/x/text/cases/cases.go generated vendored
View file

@ -1,162 +0,0 @@
// Copyright 2014 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.
//go:generate go run gen.go gen_trieval.go
// Package cases provides general and language-specific case mappers.
package cases // import "golang.org/x/text/cases"
import (
"golang.org/x/text/language"
"golang.org/x/text/transform"
)
// References:
// - Unicode Reference Manual Chapter 3.13, 4.2, and 5.18.
// - http://www.unicode.org/reports/tr29/
// - http://www.unicode.org/Public/6.3.0/ucd/CaseFolding.txt
// - http://www.unicode.org/Public/6.3.0/ucd/SpecialCasing.txt
// - http://www.unicode.org/Public/6.3.0/ucd/DerivedCoreProperties.txt
// - http://www.unicode.org/Public/6.3.0/ucd/auxiliary/WordBreakProperty.txt
// - http://www.unicode.org/Public/6.3.0/ucd/auxiliary/WordBreakTest.txt
// - http://userguide.icu-project.org/transforms/casemappings
// TODO:
// - Case folding
// - Wide and Narrow?
// - Segmenter option for title casing.
// - ASCII fast paths
// - Encode Soft-Dotted property within trie somehow.
// A Caser transforms given input to a certain case. It implements
// transform.Transformer.
//
// A Caser may be stateful and should therefore not be shared between
// goroutines.
type Caser struct {
t transform.SpanningTransformer
}
// Bytes returns a new byte slice with the result of converting b to the case
// form implemented by c.
func (c Caser) Bytes(b []byte) []byte {
b, _, _ = transform.Bytes(c.t, b)
return b
}
// String returns a string with the result of transforming s to the case form
// implemented by c.
func (c Caser) String(s string) string {
s, _, _ = transform.String(c.t, s)
return s
}
// Reset resets the Caser to be reused for new input after a previous call to
// Transform.
func (c Caser) Reset() { c.t.Reset() }
// Transform implements the transform.Transformer interface and transforms the
// given input to the case form implemented by c.
func (c Caser) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
return c.t.Transform(dst, src, atEOF)
}
// Span implements the transform.SpanningTransformer interface.
func (c Caser) Span(src []byte, atEOF bool) (n int, err error) {
return c.t.Span(src, atEOF)
}
// Upper returns a Caser for language-specific uppercasing.
func Upper(t language.Tag, opts ...Option) Caser {
return Caser{makeUpper(t, getOpts(opts...))}
}
// Lower returns a Caser for language-specific lowercasing.
func Lower(t language.Tag, opts ...Option) Caser {
return Caser{makeLower(t, getOpts(opts...))}
}
// Title returns a Caser for language-specific title casing. It uses an
// approximation of the default Unicode Word Break algorithm.
func Title(t language.Tag, opts ...Option) Caser {
return Caser{makeTitle(t, getOpts(opts...))}
}
// Fold returns a Caser that implements Unicode case folding. The returned Caser
// is stateless and safe to use concurrently by multiple goroutines.
//
// Case folding does not normalize the input and may not preserve a normal form.
// Use the collate or search package for more convenient and linguistically
// sound comparisons. Use golang.org/x/text/secure/precis for string comparisons
// where security aspects are a concern.
func Fold(opts ...Option) Caser {
return Caser{makeFold(getOpts(opts...))}
}
// An Option is used to modify the behavior of a Caser.
type Option func(o options) options
// TODO: consider these options to take a boolean as well, like FinalSigma.
// The advantage of using this approach is that other providers of a lower-case
// algorithm could set different defaults by prefixing a user-provided slice
// of options with their own. This is handy, for instance, for the precis
// package which would override the default to not handle the Greek final sigma.
var (
// NoLower disables the lowercasing of non-leading letters for a title
// caser.
NoLower Option = noLower
// Compact omits mappings in case folding for characters that would grow the
// input. (Unimplemented.)
Compact Option = compact
)
// TODO: option to preserve a normal form, if applicable?
type options struct {
noLower bool
simple bool
// TODO: segmenter, max ignorable, alternative versions, etc.
ignoreFinalSigma bool
}
func getOpts(o ...Option) (res options) {
for _, f := range o {
res = f(res)
}
return
}
func noLower(o options) options {
o.noLower = true
return o
}
func compact(o options) options {
o.simple = true
return o
}
// HandleFinalSigma specifies whether the special handling of Greek final sigma
// should be enabled. Unicode prescribes handling the Greek final sigma for all
// locales, but standards like IDNA and PRECIS override this default.
func HandleFinalSigma(enable bool) Option {
if enable {
return handleFinalSigma
}
return ignoreFinalSigma
}
func ignoreFinalSigma(o options) options {
o.ignoreFinalSigma = true
return o
}
func handleFinalSigma(o options) options {
o.ignoreFinalSigma = false
return o
}

376
vendor/golang.org/x/text/cases/context.go generated vendored
View file

@ -1,376 +0,0 @@
// Copyright 2014 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 cases
import "golang.org/x/text/transform"
// A context is used for iterating over source bytes, fetching case info and
// writing to a destination buffer.
//
// Casing operations may need more than one rune of context to decide how a rune
// should be cased. Casing implementations should call checkpoint on context
// whenever it is known to be safe to return the runes processed so far.
//
// It is recommended for implementations to not allow for more than 30 case
// ignorables as lookahead (analogous to the limit in norm) and to use state if
// unbounded lookahead is needed for cased runes.
type context struct {
dst, src []byte
atEOF bool
pDst int // pDst points past the last written rune in dst.
pSrc int // pSrc points to the start of the currently scanned rune.
// checkpoints safe to return in Transform, where nDst <= pDst and nSrc <= pSrc.
nDst, nSrc int
err error
sz int // size of current rune
info info // case information of currently scanned rune
// State preserved across calls to Transform.
isMidWord bool // false if next cased letter needs to be title-cased.
}
func (c *context) Reset() {
c.isMidWord = false
}
// ret returns the return values for the Transform method. It checks whether
// there were insufficient bytes in src to complete and introduces an error
// accordingly, if necessary.
func (c *context) ret() (nDst, nSrc int, err error) {
if c.err != nil || c.nSrc == len(c.src) {
return c.nDst, c.nSrc, c.err
}
// This point is only reached by mappers if there was no short destination
// buffer. This means that the source buffer was exhausted and that c.sz was
// set to 0 by next.
if c.atEOF && c.pSrc == len(c.src) {
return c.pDst, c.pSrc, nil
}
return c.nDst, c.nSrc, transform.ErrShortSrc
}
// retSpan returns the return values for the Span method. It checks whether
// there were insufficient bytes in src to complete and introduces an error
// accordingly, if necessary.
func (c *context) retSpan() (n int, err error) {
_, nSrc, err := c.ret()
return nSrc, err
}
// checkpoint sets the return value buffer points for Transform to the current
// positions.
func (c *context) checkpoint() {
if c.err == nil {
c.nDst, c.nSrc = c.pDst, c.pSrc+c.sz
}
}
// unreadRune causes the last rune read by next to be reread on the next
// invocation of next. Only one unreadRune may be called after a call to next.
func (c *context) unreadRune() {
c.sz = 0
}
func (c *context) next() bool {
c.pSrc += c.sz
if c.pSrc == len(c.src) || c.err != nil {
c.info, c.sz = 0, 0
return false
}
v, sz := trie.lookup(c.src[c.pSrc:])
c.info, c.sz = info(v), sz
if c.sz == 0 {
if c.atEOF {
// A zero size means we have an incomplete rune. If we are atEOF,
// this means it is an illegal rune, which we will consume one
// byte at a time.
c.sz = 1
} else {
c.err = transform.ErrShortSrc
return false
}
}
return true
}
// writeBytes adds bytes to dst.
func (c *context) writeBytes(b []byte) bool {
if len(c.dst)-c.pDst < len(b) {
c.err = transform.ErrShortDst
return false
}
// This loop is faster than using copy.
for _, ch := range b {
c.dst[c.pDst] = ch
c.pDst++
}
return true
}
// writeString writes the given string to dst.
func (c *context) writeString(s string) bool {
if len(c.dst)-c.pDst < len(s) {
c.err = transform.ErrShortDst
return false
}
// This loop is faster than using copy.
for i := 0; i < len(s); i++ {
c.dst[c.pDst] = s[i]
c.pDst++
}
return true
}
// copy writes the current rune to dst.
func (c *context) copy() bool {
return c.writeBytes(c.src[c.pSrc : c.pSrc+c.sz])
}
// copyXOR copies the current rune to dst and modifies it by applying the XOR
// pattern of the case info. It is the responsibility of the caller to ensure
// that this is a rune with a XOR pattern defined.
func (c *context) copyXOR() bool {
if !c.copy() {
return false
}
if c.info&xorIndexBit == 0 {
// Fast path for 6-bit XOR pattern, which covers most cases.
c.dst[c.pDst-1] ^= byte(c.info >> xorShift)
} else {
// Interpret XOR bits as an index.
// TODO: test performance for unrolling this loop. Verify that we have
// at least two bytes and at most three.
idx := c.info >> xorShift
for p := c.pDst - 1; ; p-- {
c.dst[p] ^= xorData[idx]
idx--
if xorData[idx] == 0 {
break
}
}
}
return true
}
// hasPrefix returns true if src[pSrc:] starts with the given string.
func (c *context) hasPrefix(s string) bool {
b := c.src[c.pSrc:]
if len(b) < len(s) {
return false
}
for i, c := range b[:len(s)] {
if c != s[i] {
return false
}
}
return true
}
// caseType returns an info with only the case bits, normalized to either
// cLower, cUpper, cTitle or cUncased.
func (c *context) caseType() info {
cm := c.info & 0x7
if cm < 4 {
return cm
}
if cm >= cXORCase {
// xor the last bit of the rune with the case type bits.
b := c.src[c.pSrc+c.sz-1]
return info(b&1) ^ cm&0x3
}
if cm == cIgnorableCased {
return cLower
}
return cUncased
}
// lower writes the lowercase version of the current rune to dst.
func lower(c *context) bool {
ct := c.caseType()
if c.info&hasMappingMask == 0 || ct == cLower {
return c.copy()
}
if c.info&exceptionBit == 0 {
return c.copyXOR()
}
e := exceptions[c.info>>exceptionShift:]
offset := 2 + e[0]&lengthMask // size of header + fold string
if nLower := (e[1] >> lengthBits) & lengthMask; nLower != noChange {
return c.writeString(e[offset : offset+nLower])
}
return c.copy()
}
func isLower(c *context) bool {
ct := c.caseType()
if c.info&hasMappingMask == 0 || ct == cLower {
return true
}
if c.info&exceptionBit == 0 {
c.err = transform.ErrEndOfSpan
return false
}
e := exceptions[c.info>>exceptionShift:]
if nLower := (e[1] >> lengthBits) & lengthMask; nLower != noChange {
c.err = transform.ErrEndOfSpan
return false
}
return true
}
// upper writes the uppercase version of the current rune to dst.
func upper(c *context) bool {
ct := c.caseType()
if c.info&hasMappingMask == 0 || ct == cUpper {
return c.copy()
}
if c.info&exceptionBit == 0 {
return c.copyXOR()
}
e := exceptions[c.info>>exceptionShift:]
offset := 2 + e[0]&lengthMask // size of header + fold string
// Get length of first special case mapping.
n := (e[1] >> lengthBits) & lengthMask
if ct == cTitle {
// The first special case mapping is for lower. Set n to the second.
if n == noChange {
n = 0
}
n, e = e[1]&lengthMask, e[n:]
}
if n != noChange {
return c.writeString(e[offset : offset+n])
}
return c.copy()
}
// isUpper writes the isUppercase version of the current rune to dst.
func isUpper(c *context) bool {
ct := c.caseType()
if c.info&hasMappingMask == 0 || ct == cUpper {
return true
}
if c.info&exceptionBit == 0 {
c.err = transform.ErrEndOfSpan
return false
}
e := exceptions[c.info>>exceptionShift:]
// Get length of first special case mapping.
n := (e[1] >> lengthBits) & lengthMask
if ct == cTitle {
n = e[1] & lengthMask
}
if n != noChange {
c.err = transform.ErrEndOfSpan
return false
}
return true
}
// title writes the title case version of the current rune to dst.
func title(c *context) bool {
ct := c.caseType()
if c.info&hasMappingMask == 0 || ct == cTitle {
return c.copy()
}
if c.info&exceptionBit == 0 {
if ct == cLower {
return c.copyXOR()
}
return c.copy()
}
// Get the exception data.
e := exceptions[c.info>>exceptionShift:]
offset := 2 + e[0]&lengthMask // size of header + fold string
nFirst := (e[1] >> lengthBits) & lengthMask
if nTitle := e[1] & lengthMask; nTitle != noChange {
if nFirst != noChange {
e = e[nFirst:]
}
return c.writeString(e[offset : offset+nTitle])
}
if ct == cLower && nFirst != noChange {
// Use the uppercase version instead.
return c.writeString(e[offset : offset+nFirst])
}
// Already in correct case.
return c.copy()
}
// isTitle reports whether the current rune is in title case.
func isTitle(c *context) bool {
ct := c.caseType()
if c.info&hasMappingMask == 0 || ct == cTitle {
return true
}
if c.info&exceptionBit == 0 {
if ct == cLower {
c.err = transform.ErrEndOfSpan
return false
}
return true
}
// Get the exception data.
e := exceptions[c.info>>exceptionShift:]
if nTitle := e[1] & lengthMask; nTitle != noChange {
c.err = transform.ErrEndOfSpan
return false
}
nFirst := (e[1] >> lengthBits) & lengthMask
if ct == cLower && nFirst != noChange {
c.err = transform.ErrEndOfSpan
return false
}
return true
}
// foldFull writes the foldFull version of the current rune to dst.
func foldFull(c *context) bool {
if c.info&hasMappingMask == 0 {
return c.copy()
}
ct := c.caseType()
if c.info&exceptionBit == 0 {
if ct != cLower || c.info&inverseFoldBit != 0 {
return c.copyXOR()
}
return c.copy()
}
e := exceptions[c.info>>exceptionShift:]
n := e[0] & lengthMask
if n == 0 {
if ct == cLower {
return c.copy()
}
n = (e[1] >> lengthBits) & lengthMask
}
return c.writeString(e[2 : 2+n])
}
// isFoldFull reports whether the current run is mapped to foldFull
func isFoldFull(c *context) bool {
if c.info&hasMappingMask == 0 {
return true
}
ct := c.caseType()
if c.info&exceptionBit == 0 {
if ct != cLower || c.info&inverseFoldBit != 0 {
c.err = transform.ErrEndOfSpan
return false
}
return true
}
e := exceptions[c.info>>exceptionShift:]
n := e[0] & lengthMask
if n == 0 && ct == cLower {
return true
}
c.err = transform.ErrEndOfSpan
return false
}

34
vendor/golang.org/x/text/cases/fold.go generated vendored
View file

@ -1,34 +0,0 @@
// Copyright 2016 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 cases
import "golang.org/x/text/transform"
type caseFolder struct{ transform.NopResetter }
// caseFolder implements the Transformer interface for doing case folding.
func (t *caseFolder) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
c := context{dst: dst, src: src, atEOF: atEOF}
for c.next() {
foldFull(&c)
c.checkpoint()
}
return c.ret()
}
func (t *caseFolder) Span(src []byte, atEOF bool) (n int, err error) {
c := context{src: src, atEOF: atEOF}
for c.next() && isFoldFull(&c) {
c.checkpoint()
}
return c.retSpan()
}
func makeFold(o options) transform.SpanningTransformer {
// TODO: Special case folding, through option Language, Special/Turkic, or
// both.
// TODO: Implement Compact options.
return &caseFolder{}
}

839
vendor/golang.org/x/text/cases/gen.go generated vendored
View file

@ -1,839 +0,0 @@
// Copyright 2014 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
// This program generates the trie for casing operations. The Unicode casing
// algorithm requires the lookup of various properties and mappings for each
// rune. The table generated by this generator combines several of the most
// frequently used of these into a single trie so that they can be accessed
// with a single lookup.
package main
import (
"bytes"
"fmt"
"io"
"io/ioutil"
"log"
"reflect"
"strconv"
"strings"
"unicode"
"golang.org/x/text/internal/gen"
"golang.org/x/text/internal/triegen"
"golang.org/x/text/internal/ucd"
"golang.org/x/text/unicode/norm"
)
func main() {
gen.Init()
genTables()
genTablesTest()
gen.Repackage("gen_trieval.go", "trieval.go", "cases")
}
// runeInfo contains all information for a rune that we care about for casing
// operations.
type runeInfo struct {
Rune rune
entry info // trie value for this rune.
CaseMode info
// Simple case mappings.
Simple [1 + maxCaseMode][]rune
// Special casing
HasSpecial bool
Conditional bool
Special [1 + maxCaseMode][]rune
// Folding
FoldSimple rune
FoldSpecial rune
FoldFull []rune
// TODO: FC_NFKC, or equivalent data.
// Properties
SoftDotted bool
CaseIgnorable bool
Cased bool
DecomposeGreek bool
BreakType string
BreakCat breakCategory
// We care mostly about 0, Above, and IotaSubscript.
CCC byte
}
type breakCategory int
const (
breakBreak breakCategory = iota
breakLetter
breakMid
)
// mapping returns the case mapping for the given case type.
func (r *runeInfo) mapping(c info) string {
if r.HasSpecial {
return string(r.Special[c])
}
if len(r.Simple[c]) != 0 {
return string(r.Simple[c])
}
return string(r.Rune)
}
func parse(file string, f func(p *ucd.Parser)) {
ucd.Parse(gen.OpenUCDFile(file), f)
}
func parseUCD() []runeInfo {
chars := make([]runeInfo, unicode.MaxRune)
get := func(r rune) *runeInfo {
c := &chars[r]
c.Rune = r
return c
}
parse("UnicodeData.txt", func(p *ucd.Parser) {
ri := get(p.Rune(0))
ri.CCC = byte(p.Int(ucd.CanonicalCombiningClass))
ri.Simple[cLower] = p.Runes(ucd.SimpleLowercaseMapping)
ri.Simple[cUpper] = p.Runes(ucd.SimpleUppercaseMapping)
ri.Simple[cTitle] = p.Runes(ucd.SimpleTitlecaseMapping)
if p.String(ucd.GeneralCategory) == "Lt" {
ri.CaseMode = cTitle
}
})
// <code>; <property>
parse("PropList.txt", func(p *ucd.Parser) {
if p.String(1) == "Soft_Dotted" {
chars[p.Rune(0)].SoftDotted = true
}
})
// <code>; <word break type>
parse("DerivedCoreProperties.txt", func(p *ucd.Parser) {
ri := get(p.Rune(0))
switch p.String(1) {
case "Case_Ignorable":
ri.CaseIgnorable = true
case "Cased":
ri.Cased = true
case "Lowercase":
ri.CaseMode = cLower
case "Uppercase":
ri.CaseMode = cUpper
}
})
// <code>; <lower> ; <title> ; <upper> ; (<condition_list> ;)?
parse("SpecialCasing.txt", func(p *ucd.Parser) {
// We drop all conditional special casing and deal with them manually in
// the language-specific case mappers. Rune 0x03A3 is the only one with
// a conditional formatting that is not language-specific. However,
// dealing with this letter is tricky, especially in a streaming
// context, so we deal with it in the Caser for Greek specifically.
ri := get(p.Rune(0))
if p.String(4) == "" {
ri.HasSpecial = true
ri.Special[cLower] = p.Runes(1)
ri.Special[cTitle] = p.Runes(2)
ri.Special[cUpper] = p.Runes(3)
} else {
ri.Conditional = true
}
})
// TODO: Use text breaking according to UAX #29.
// <code>; <word break type>
parse("auxiliary/WordBreakProperty.txt", func(p *ucd.Parser) {
ri := get(p.Rune(0))
ri.BreakType = p.String(1)
// We collapse the word breaking properties onto the categories we need.
switch p.String(1) { // TODO: officially we need to canonicalize.
case "MidLetter", "MidNumLet", "Single_Quote":
ri.BreakCat = breakMid
if !ri.CaseIgnorable {
// finalSigma relies on the fact that all breakMid runes are
// also a Case_Ignorable. Revisit this code when this changes.
log.Fatalf("Rune %U, which has a break category mid, is not a case ignorable", ri)
}
case "ALetter", "Hebrew_Letter", "Numeric", "Extend", "ExtendNumLet", "Format", "ZWJ":
ri.BreakCat = breakLetter
}
})
// <code>; <type>; <mapping>
parse("CaseFolding.txt", func(p *ucd.Parser) {
ri := get(p.Rune(0))
switch p.String(1) {
case "C":
ri.FoldSimple = p.Rune(2)
ri.FoldFull = p.Runes(2)
case "S":
ri.FoldSimple = p.Rune(2)
case "T":
ri.FoldSpecial = p.Rune(2)
case "F":
ri.FoldFull = p.Runes(2)
default:
log.Fatalf("%U: unknown type: %s", p.Rune(0), p.String(1))
}
})
return chars
}
func genTables() {
chars := parseUCD()
verifyProperties(chars)
t := triegen.NewTrie("case")
for i := range chars {
c := &chars[i]
makeEntry(c)
t.Insert(rune(i), uint64(c.entry))
}
w := gen.NewCodeWriter()
defer w.WriteGoFile("tables.go", "cases")
gen.WriteUnicodeVersion(w)
// TODO: write CLDR version after adding a mechanism to detect that the
// tables on which the manually created locale-sensitive casing code is
// based hasn't changed.
w.WriteVar("xorData", string(xorData))
w.WriteVar("exceptions", string(exceptionData))
sz, err := t.Gen(w, triegen.Compact(&sparseCompacter{}))
if err != nil {
log.Fatal(err)
}
w.Size += sz
}
func makeEntry(ri *runeInfo) {
if ri.CaseIgnorable {
if ri.Cased {
ri.entry = cIgnorableCased
} else {
ri.entry = cIgnorableUncased
}
} else {
ri.entry = ri.CaseMode
}
// TODO: handle soft-dotted.
ccc := cccOther
switch ri.CCC {
case 0: // Not_Reordered
ccc = cccZero
case above: // Above
ccc = cccAbove
}
switch ri.BreakCat {
case breakBreak:
ccc = cccBreak
case breakMid:
ri.entry |= isMidBit
}
ri.entry |= ccc
if ri.CaseMode == cUncased {
return
}
// Need to do something special.
if ri.CaseMode == cTitle || ri.HasSpecial || ri.mapping(cTitle) != ri.mapping(cUpper) {
makeException(ri)
return
}
if f := string(ri.FoldFull); len(f) > 0 && f != ri.mapping(cUpper) && f != ri.mapping(cLower) {
makeException(ri)
return
}
// Rune is either lowercase or uppercase.
orig := string(ri.Rune)
mapped := ""
if ri.CaseMode == cUpper {
mapped = ri.mapping(cLower)
} else {
mapped = ri.mapping(cUpper)
}
if len(orig) != len(mapped) {
makeException(ri)
return
}
if string(ri.FoldFull) == ri.mapping(cUpper) {
ri.entry |= inverseFoldBit
}
n := len(orig)
// Create per-byte XOR mask.
var b []byte
for i := 0; i < n; i++ {
b = append(b, orig[i]^mapped[i])
}
// Remove leading 0 bytes, but keep at least one byte.
for ; len(b) > 1 && b[0] == 0; b = b[1:] {
}
if len(b) == 1 && b[0]&0xc0 == 0 {
ri.entry |= info(b[0]) << xorShift
return
}
key := string(b)
x, ok := xorCache[key]
if !ok {
xorData = append(xorData, 0) // for detecting start of sequence
xorData = append(xorData, b...)
x = len(xorData) - 1
xorCache[key] = x
}
ri.entry |= info(x<<xorShift) | xorIndexBit
}
var xorCache = map[string]int{}
// xorData contains byte-wise XOR data for the least significant bytes of a
// UTF-8 encoded rune. An index points to the last byte. The sequence starts
// with a zero terminator.
var xorData = []byte{}
// See the comments in gen_trieval.go re "the exceptions slice".
var exceptionData = []byte{0}
// makeException encodes case mappings that cannot be expressed in a simple
// XOR diff.
func makeException(ri *runeInfo) {
ccc := ri.entry & cccMask
// Set exception bit and retain case type.
ri.entry &= 0x0007
ri.entry |= exceptionBit
if len(exceptionData) >= 1<<numExceptionBits {
log.Fatalf("%U:exceptionData too large %x > %d bits", ri.Rune, len(exceptionData), numExceptionBits)
}
// Set the offset in the exceptionData array.
ri.entry |= info(len(exceptionData) << exceptionShift)
orig := string(ri.Rune)
tc := ri.mapping(cTitle)
uc := ri.mapping(cUpper)
lc := ri.mapping(cLower)
ff := string(ri.FoldFull)
// addString sets the length of a string and adds it to the expansions array.
addString := func(s string, b *byte) {
if len(s) == 0 {
// Zero-length mappings exist, but only for conditional casing,
// which we are representing outside of this table.
log.Fatalf("%U: has zero-length mapping.", ri.Rune)
}
*b <<= 3
if s != orig {
n := len(s)
if n > 7 {
log.Fatalf("%U: mapping larger than 7 (%d)", ri.Rune, n)
}
*b |= byte(n)
exceptionData = append(exceptionData, s...)
}
}
// byte 0:
exceptionData = append(exceptionData, byte(ccc)|byte(len(ff)))
// byte 1:
p := len(exceptionData)
exceptionData = append(exceptionData, 0)
if len(ff) > 7 { // May be zero-length.
log.Fatalf("%U: fold string larger than 7 (%d)", ri.Rune, len(ff))
}
exceptionData = append(exceptionData, ff...)
ct := ri.CaseMode
if ct != cLower {
addString(lc, &exceptionData[p])
}
if ct != cUpper {
addString(uc, &exceptionData[p])
}
if ct != cTitle {
// If title is the same as upper, we set it to the original string so
// that it will be marked as not present. This implies title case is
// the same as upper case.
if tc == uc {
tc = orig
}
addString(tc, &exceptionData[p])
}
}
// sparseCompacter is a trie value block Compacter. There are many cases where
// successive runes alternate between lower- and upper-case. This Compacter
// exploits this by adding a special case type where the case value is obtained
// from or-ing it with the least-significant bit of the rune, creating large
// ranges of equal case values that compress well.
type sparseCompacter struct {
sparseBlocks [][]uint16
sparseOffsets []uint16
sparseCount int
}
// makeSparse returns the number of elements that compact block would contain
// as well as the modified values.
func makeSparse(vals []uint64) ([]uint16, int) {
// Copy the values.
values := make([]uint16, len(vals))
for i, v := range vals {
values[i] = uint16(v)
}
alt := func(i int, v uint16) uint16 {
if cm := info(v & fullCasedMask); cm == cUpper || cm == cLower {
// Convert cLower or cUpper to cXORCase value, which has the form 11x.
xor := v
xor &^= 1
xor |= uint16(i&1) ^ (v & 1)
xor |= 0x4
return xor
}
return v
}
var count int
var previous uint16
for i, v := range values {
if v != 0 {
// Try if the unmodified value is equal to the previous.
if v == previous {
continue
}
// Try if the xor-ed value is equal to the previous value.
a := alt(i, v)
if a == previous {
values[i] = a
continue
}
// This is a new value.
count++
// Use the xor-ed value if it will be identical to the next value.
if p := i + 1; p < len(values) && alt(p, values[p]) == a {
values[i] = a
v = a
}
}
previous = v
}
return values, count
}
func (s *sparseCompacter) Size(v []uint64) (int, bool) {
_, n := makeSparse(v)
// We limit using this method to having 16 entries.
if n > 16 {
return 0, false
}
return 2 + int(reflect.TypeOf(valueRange{}).Size())*n, true
}
func (s *sparseCompacter) Store(v []uint64) uint32 {
h := uint32(len(s.sparseOffsets))
values, sz := makeSparse(v)
s.sparseBlocks = append(s.sparseBlocks, values)
s.sparseOffsets = append(s.sparseOffsets, uint16(s.sparseCount))
s.sparseCount += sz
return h
}
func (s *sparseCompacter) Handler() string {
// The sparse global variable and its lookup method is defined in gen_trieval.go.
return "sparse.lookup"
}
func (s *sparseCompacter) Print(w io.Writer) (retErr error) {
p := func(format string, args ...interface{}) {
_, err := fmt.Fprintf(w, format, args...)
if retErr == nil && err != nil {
retErr = err
}
}
ls := len(s.sparseBlocks)
if ls == len(s.sparseOffsets) {
s.sparseOffsets = append(s.sparseOffsets, uint16(s.sparseCount))
}
p("// sparseOffsets: %d entries, %d bytes\n", ls+1, (ls+1)*2)
p("var sparseOffsets = %#v\n\n", s.sparseOffsets)
ns := s.sparseCount
p("// sparseValues: %d entries, %d bytes\n", ns, ns*4)
p("var sparseValues = [%d]valueRange {", ns)
for i, values := range s.sparseBlocks {
p("\n// Block %#x, offset %#x", i, s.sparseOffsets[i])
var v uint16
for i, nv := range values {
if nv != v {
if v != 0 {
p(",hi:%#02x},", 0x80+i-1)
}
if nv != 0 {
p("\n{value:%#04x,lo:%#02x", nv, 0x80+i)
}
}
v = nv
}
if v != 0 {
p(",hi:%#02x},", 0x80+len(values)-1)
}
}
p("\n}\n\n")
return
}
// verifyProperties that properties of the runes that are relied upon in the
// implementation. Each property is marked with an identifier that is referred
// to in the places where it is used.
func verifyProperties(chars []runeInfo) {
for i, c := range chars {
r := rune(i)
// Rune properties.
// A.1: modifier never changes on lowercase. [ltLower]
if c.CCC > 0 && unicode.ToLower(r) != r {
log.Fatalf("%U: non-starter changes when lowercased", r)
}
// A.2: properties of decompositions starting with I or J. [ltLower]
d := norm.NFD.PropertiesString(string(r)).Decomposition()
if len(d) > 0 {
if d[0] == 'I' || d[0] == 'J' {
// A.2.1: we expect at least an ASCII character and a modifier.
if len(d) < 3 {
log.Fatalf("%U: length of decomposition was %d; want >= 3", r, len(d))
}
// All subsequent runes are modifiers and all have the same CCC.
runes := []rune(string(d[1:]))
ccc := chars[runes[0]].CCC
for _, mr := range runes[1:] {
mc := chars[mr]
// A.2.2: all modifiers have a CCC of Above or less.
if ccc == 0 || ccc > above {
log.Fatalf("%U: CCC of successive rune (%U) was %d; want (0,230]", r, mr, ccc)
}
// A.2.3: a sequence of modifiers all have the same CCC.
if mc.CCC != ccc {
log.Fatalf("%U: CCC of follow-up modifier (%U) was %d; want %d", r, mr, mc.CCC, ccc)
}
// A.2.4: for each trailing r, r in [0x300, 0x311] <=> CCC == Above.
if (ccc == above) != (0x300 <= mr && mr <= 0x311) {
log.Fatalf("%U: modifier %U in [U+0300, U+0311] != ccc(%U) == 230", r, mr, mr)
}
if i += len(string(mr)); i >= len(d) {
break
}
}
}
}
// A.3: no U+0307 in decomposition of Soft-Dotted rune. [ltUpper]
if unicode.Is(unicode.Soft_Dotted, r) && strings.Contains(string(d), "\u0307") {
log.Fatalf("%U: decomposition of soft-dotted rune may not contain U+0307", r)
}
// A.4: only rune U+0345 may be of CCC Iota_Subscript. [elUpper]
if c.CCC == iotaSubscript && r != 0x0345 {
log.Fatalf("%U: only rune U+0345 may have CCC Iota_Subscript", r)
}
// A.5: soft-dotted runes do not have exceptions.
if c.SoftDotted && c.entry&exceptionBit != 0 {
log.Fatalf("%U: soft-dotted has exception", r)
}
// A.6: Greek decomposition. [elUpper]
if unicode.Is(unicode.Greek, r) {
if b := norm.NFD.PropertiesString(string(r)).Decomposition(); b != nil {
runes := []rune(string(b))
// A.6.1: If a Greek rune decomposes and the first rune of the
// decomposition is greater than U+00FF, the rune is always
// great and not a modifier.
if f := runes[0]; unicode.IsMark(f) || f > 0xFF && !unicode.Is(unicode.Greek, f) {
log.Fatalf("%U: expeced first rune of Greek decomposition to be letter, found %U", r, f)
}
// A.6.2: Any follow-up rune in a Greek decomposition is a
// modifier of which the first should be gobbled in
// decomposition.
for _, m := range runes[1:] {
switch m {
case 0x0313, 0x0314, 0x0301, 0x0300, 0x0306, 0x0342, 0x0308, 0x0304, 0x345:
default:
log.Fatalf("%U: modifier %U is outside of expeced Greek modifier set", r, m)
}
}
}
}
// Breaking properties.
// B.1: all runes with CCC > 0 are of break type Extend.
if c.CCC > 0 && c.BreakType != "Extend" {
log.Fatalf("%U: CCC == %d, but got break type %s; want Extend", r, c.CCC, c.BreakType)
}
// B.2: all cased runes with c.CCC == 0 are of break type ALetter.
if c.CCC == 0 && c.Cased && c.BreakType != "ALetter" {
log.Fatalf("%U: cased, but got break type %s; want ALetter", r, c.BreakType)
}
// B.3: letter category.
if c.CCC == 0 && c.BreakCat != breakBreak && !c.CaseIgnorable {
if c.BreakCat != breakLetter {
log.Fatalf("%U: check for letter break type gave %d; want %d", r, c.BreakCat, breakLetter)
}
}
}
}
func genTablesTest() {
w := &bytes.Buffer{}
fmt.Fprintln(w, "var (")
printProperties(w, "DerivedCoreProperties.txt", "Case_Ignorable", verifyIgnore)
// We discard the output as we know we have perfect functions. We run them
// just to verify the properties are correct.
n := printProperties(ioutil.Discard, "DerivedCoreProperties.txt", "Cased", verifyCased)
n += printProperties(ioutil.Discard, "DerivedCoreProperties.txt", "Lowercase", verifyLower)
n += printProperties(ioutil.Discard, "DerivedCoreProperties.txt", "Uppercase", verifyUpper)
if n > 0 {
log.Fatalf("One of the discarded properties does not have a perfect filter.")
}
// <code>; <lower> ; <title> ; <upper> ; (<condition_list> ;)?
fmt.Fprintln(w, "\tspecial = map[rune]struct{ toLower, toTitle, toUpper string }{")
parse("SpecialCasing.txt", func(p *ucd.Parser) {
// Skip conditional entries.
if p.String(4) != "" {
return
}
r := p.Rune(0)
fmt.Fprintf(w, "\t\t0x%04x: {%q, %q, %q},\n",
r, string(p.Runes(1)), string(p.Runes(2)), string(p.Runes(3)))
})
fmt.Fprint(w, "\t}\n\n")
// <code>; <type>; <runes>
table := map[rune]struct{ simple, full, special string }{}
parse("CaseFolding.txt", func(p *ucd.Parser) {
r := p.Rune(0)
t := p.String(1)
v := string(p.Runes(2))
if t != "T" && v == string(unicode.ToLower(r)) {
return
}
x := table[r]
switch t {
case "C":
x.full = v
x.simple = v
case "S":
x.simple = v
case "F":
x.full = v
case "T":
x.special = v
}
table[r] = x
})
fmt.Fprintln(w, "\tfoldMap = map[rune]struct{ simple, full, special string }{")
for r := rune(0); r < 0x10FFFF; r++ {
x, ok := table[r]
if !ok {
continue
}
fmt.Fprintf(w, "\t\t0x%04x: {%q, %q, %q},\n", r, x.simple, x.full, x.special)
}
fmt.Fprint(w, "\t}\n\n")
// Break property
notBreak := map[rune]bool{}
parse("auxiliary/WordBreakProperty.txt", func(p *ucd.Parser) {
switch p.String(1) {
case "Extend", "Format", "MidLetter", "MidNumLet", "Single_Quote",
"ALetter", "Hebrew_Letter", "Numeric", "ExtendNumLet", "ZWJ":
notBreak[p.Rune(0)] = true
}
})
fmt.Fprintln(w, "\tbreakProp = []struct{ lo, hi rune }{")
inBreak := false
for r := rune(0); r <= lastRuneForTesting; r++ {
if isBreak := !notBreak[r]; isBreak != inBreak {
if isBreak {
fmt.Fprintf(w, "\t\t{0x%x, ", r)
} else {
fmt.Fprintf(w, "0x%x},\n", r-1)
}
inBreak = isBreak
}
}
if inBreak {
fmt.Fprintf(w, "0x%x},\n", lastRuneForTesting)
}
fmt.Fprint(w, "\t}\n\n")
// Word break test
// Filter out all samples that do not contain cased characters.
cased := map[rune]bool{}
parse("DerivedCoreProperties.txt", func(p *ucd.Parser) {
if p.String(1) == "Cased" {
cased[p.Rune(0)] = true
}
})
fmt.Fprintln(w, "\tbreakTest = []string{")
parse("auxiliary/WordBreakTest.txt", func(p *ucd.Parser) {
c := strings.Split(p.String(0), " ")
const sep = '|'
numCased := 0
test := ""
for ; len(c) >= 2; c = c[2:] {
if c[0] == "÷" && test != "" {
test += string(sep)
}
i, err := strconv.ParseUint(c[1], 16, 32)
r := rune(i)
if err != nil {
log.Fatalf("Invalid rune %q.", c[1])
}
if r == sep {
log.Fatalf("Separator %q not allowed in test data. Pick another one.", sep)
}
if cased[r] {
numCased++
}
test += string(r)
}
if numCased > 1 {
fmt.Fprintf(w, "\t\t%q,\n", test)
}
})
fmt.Fprintln(w, "\t}")
fmt.Fprintln(w, ")")
gen.WriteGoFile("tables_test.go", "cases", w.Bytes())
}
// These functions are just used for verification that their definition have not
// changed in the Unicode Standard.
func verifyCased(r rune) bool {
return verifyLower(r) || verifyUpper(r) || unicode.IsTitle(r)
}
func verifyLower(r rune) bool {
return unicode.IsLower(r) || unicode.Is(unicode.Other_Lowercase, r)
}
func verifyUpper(r rune) bool {
return unicode.IsUpper(r) || unicode.Is(unicode.Other_Uppercase, r)
}
// verifyIgnore is an approximation of the Case_Ignorable property using the
// core unicode package. It is used to reduce the size of the test data.
func verifyIgnore(r rune) bool {
props := []*unicode.RangeTable{
unicode.Mn,
unicode.Me,
unicode.Cf,
unicode.Lm,
unicode.Sk,
}
for _, p := range props {
if unicode.Is(p, r) {
return true
}
}
return false
}
// printProperties prints tables of rune properties from the given UCD file.
// A filter func f can be given to exclude certain values. A rune r will have
// the indicated property if it is in the generated table or if f(r).
func printProperties(w io.Writer, file, property string, f func(r rune) bool) int {
verify := map[rune]bool{}
n := 0
varNameParts := strings.Split(property, "_")
varNameParts[0] = strings.ToLower(varNameParts[0])
fmt.Fprintf(w, "\t%s = map[rune]bool{\n", strings.Join(varNameParts, ""))
parse(file, func(p *ucd.Parser) {
if p.String(1) == property {
r := p.Rune(0)
verify[r] = true
if !f(r) {
n++
fmt.Fprintf(w, "\t\t0x%.4x: true,\n", r)
}
}
})
fmt.Fprint(w, "\t}\n\n")
// Verify that f is correct, that is, it represents a subset of the property.
for r := rune(0); r <= lastRuneForTesting; r++ {
if !verify[r] && f(r) {
log.Fatalf("Incorrect filter func for property %q.", property)
}
}
return n
}
// The newCaseTrie, sparseValues and sparseOffsets definitions below are
// placeholders referred to by gen_trieval.go. The real definitions are
// generated by this program and written to tables.go.
func newCaseTrie(int) int { return 0 }
var (
sparseValues [0]valueRange
sparseOffsets [0]uint16
)

219
vendor/golang.org/x/text/cases/gen_trieval.go generated vendored
View file

@ -1,219 +0,0 @@
// Copyright 2014 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 contains definitions for interpreting the trie value of the case
// trie generated by "go run gen*.go". It is shared by both the generator
// program and the resultant package. Sharing is achieved by the generator
// copying gen_trieval.go to trieval.go and changing what's above this comment.
// info holds case information for a single rune. It is the value returned
// by a trie lookup. Most mapping information can be stored in a single 16-bit
// value. If not, for example when a rune is mapped to multiple runes, the value
// stores some basic case data and an index into an array with additional data.
//
// The per-rune values have the following format:
//
// if (exception) {
// 15..5 unsigned exception index
// 4 unused
// } else {
// 15..8 XOR pattern or index to XOR pattern for case mapping
// Only 13..8 are used for XOR patterns.
// 7 inverseFold (fold to upper, not to lower)
// 6 index: interpret the XOR pattern as an index
// or isMid if case mode is cIgnorableUncased.
// 5..4 CCC: zero (normal or break), above or other
// }
// 3 exception: interpret this value as an exception index
// (TODO: is this bit necessary? Probably implied from case mode.)
// 2..0 case mode
//
// For the non-exceptional cases, a rune must be either uncased, lowercase or
// uppercase. If the rune is cased, the XOR pattern maps either a lowercase
// rune to uppercase or an uppercase rune to lowercase (applied to the 10
// least-significant bits of the rune).
//
// See the definitions below for a more detailed description of the various
// bits.
type info uint16
const (
casedMask = 0x0003
fullCasedMask = 0x0007
ignorableMask = 0x0006
ignorableValue = 0x0004
inverseFoldBit = 1 << 7
isMidBit = 1 << 6
exceptionBit = 1 << 3
exceptionShift = 5
numExceptionBits = 11
xorIndexBit = 1 << 6
xorShift = 8
// There is no mapping if all xor bits and the exception bit are zero.
hasMappingMask = 0xff80 | exceptionBit
)
// The case mode bits encodes the case type of a rune. This includes uncased,
// title, upper and lower case and case ignorable. (For a definition of these
// terms see Chapter 3 of The Unicode Standard Core Specification.) In some rare
// cases, a rune can be both cased and case-ignorable. This is encoded by
// cIgnorableCased. A rune of this type is always lower case. Some runes are
// cased while not having a mapping.
//
// A common pattern for scripts in the Unicode standard is for upper and lower
// case runes to alternate for increasing rune values (e.g. the accented Latin
// ranges starting from U+0100 and U+1E00 among others and some Cyrillic
// characters). We use this property by defining a cXORCase mode, where the case
// mode (always upper or lower case) is derived from the rune value. As the XOR
// pattern for case mappings is often identical for successive runes, using
// cXORCase can result in large series of identical trie values. This, in turn,
// allows us to better compress the trie blocks.
const (
cUncased info = iota // 000
cTitle // 001
cLower // 010
cUpper // 011
cIgnorableUncased // 100
cIgnorableCased // 101 // lower case if mappings exist
cXORCase // 11x // case is cLower | ((rune&1) ^ x)
maxCaseMode = cUpper
)
func (c info) isCased() bool {
return c&casedMask != 0
}
func (c info) isCaseIgnorable() bool {
return c&ignorableMask == ignorableValue
}
func (c info) isNotCasedAndNotCaseIgnorable() bool {
return c&fullCasedMask == 0
}
func (c info) isCaseIgnorableAndNotCased() bool {
return c&fullCasedMask == cIgnorableUncased
}
func (c info) isMid() bool {
return c&(fullCasedMask|isMidBit) == isMidBit|cIgnorableUncased
}
// The case mapping implementation will need to know about various Canonical
// Combining Class (CCC) values. We encode two of these in the trie value:
// cccZero (0) and cccAbove (230). If the value is cccOther, it means that
// CCC(r) > 0, but not 230. A value of cccBreak means that CCC(r) == 0 and that
// the rune also has the break category Break (see below).
const (
cccBreak info = iota << 4
cccZero
cccAbove
cccOther
cccMask = cccBreak | cccZero | cccAbove | cccOther
)
const (
starter = 0
above = 230
iotaSubscript = 240
)
// The exceptions slice holds data that does not fit in a normal info entry.
// The entry is pointed to by the exception index in an entry. It has the
// following format:
//
// Header
// byte 0:
// 7..6 unused
// 5..4 CCC type (same bits as entry)
// 3 unused
// 2..0 length of fold
//
// byte 1:
// 7..6 unused
// 5..3 length of 1st mapping of case type
// 2..0 length of 2nd mapping of case type
//
// case 1st 2nd
// lower -> upper, title
// upper -> lower, title
// title -> lower, upper
//
// Lengths with the value 0x7 indicate no value and implies no change.
// A length of 0 indicates a mapping to zero-length string.
//
// Body bytes:
// case folding bytes
// lowercase mapping bytes
// uppercase mapping bytes
// titlecase mapping bytes
// closure mapping bytes (for NFKC_Casefold). (TODO)
//
// Fallbacks:
// missing fold -> lower
// missing title -> upper
// all missing -> original rune
//
// exceptions starts with a dummy byte to enforce that there is no zero index
// value.
const (
lengthMask = 0x07
lengthBits = 3
noChange = 0
)
// References to generated trie.
var trie = newCaseTrie(0)
var sparse = sparseBlocks{
values: sparseValues[:],
offsets: sparseOffsets[:],
}
// Sparse block lookup code.
// valueRange is an entry in a sparse block.
type valueRange struct {
value uint16
lo, hi byte
}
type sparseBlocks struct {
values []valueRange
offsets []uint16
}
// lookup returns the value from values block n for byte b using binary search.
func (s *sparseBlocks) lookup(n uint32, b byte) uint16 {
lo := s.offsets[n]
hi := s.offsets[n+1]
for lo < hi {
m := lo + (hi-lo)/2
r := s.values[m]
if r.lo <= b && b <= r.hi {
return r.value
}
if b < r.lo {
hi = m
} else {
lo = m + 1
}
}
return 0
}
// lastRuneForTesting is the last rune used for testing. Everything after this
// is boring.
const lastRuneForTesting = rune(0x1FFFF)

61
vendor/golang.org/x/text/cases/icu.go generated vendored
View file

@ -1,61 +0,0 @@
// Copyright 2016 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 icu
package cases
// Ideally these functions would be defined in a test file, but go test doesn't
// allow CGO in tests. The build tag should ensure either way that these
// functions will not end up in the package.
// TODO: Ensure that the correct ICU version is set.
/*
#cgo LDFLAGS: -licui18n.57 -licuuc.57
#include <stdlib.h>
#include <unicode/ustring.h>
#include <unicode/utypes.h>
#include <unicode/localpointer.h>
#include <unicode/ucasemap.h>
*/
import "C"
import "unsafe"
func doICU(tag, caser, input string) string {
err := C.UErrorCode(0)
loc := C.CString(tag)
cm := C.ucasemap_open(loc, C.uint32_t(0), &err)
buf := make([]byte, len(input)*4)
dst := (*C.char)(unsafe.Pointer(&buf[0]))
src := C.CString(input)
cn := C.int32_t(0)
switch caser {
case "fold":
cn = C.ucasemap_utf8FoldCase(cm,
dst, C.int32_t(len(buf)),
src, C.int32_t(len(input)),
&err)
case "lower":
cn = C.ucasemap_utf8ToLower(cm,
dst, C.int32_t(len(buf)),
src, C.int32_t(len(input)),
&err)
case "upper":
cn = C.ucasemap_utf8ToUpper(cm,
dst, C.int32_t(len(buf)),
src, C.int32_t(len(input)),
&err)
case "title":
cn = C.ucasemap_utf8ToTitle(cm,
dst, C.int32_t(len(buf)),
src, C.int32_t(len(input)),
&err)
}
return string(buf[:cn])
}

82
vendor/golang.org/x/text/cases/info.go generated vendored
View file

@ -1,82 +0,0 @@
// 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.
package cases
func (c info) cccVal() info {
if c&exceptionBit != 0 {
return info(exceptions[c>>exceptionShift]) & cccMask
}
return c & cccMask
}
func (c info) cccType() info {
ccc := c.cccVal()
if ccc <= cccZero {
return cccZero
}
return ccc
}
// TODO: Implement full Unicode breaking algorithm:
// 1) Implement breaking in separate package.
// 2) Use the breaker here.
// 3) Compare table size and performance of using the more generic breaker.
//
// Note that we can extend the current algorithm to be much more accurate. This
// only makes sense, though, if the performance and/or space penalty of using
// the generic breaker is big. Extra data will only be needed for non-cased
// runes, which means there are sufficient bits left in the caseType.
// ICU prohibits breaking in such cases as well.
// For the purpose of title casing we use an approximation of the Unicode Word
// Breaking algorithm defined in Annex #29:
// http://www.unicode.org/reports/tr29/#Default_Grapheme_Cluster_Table.
//
// For our approximation, we group the Word Break types into the following
// categories, with associated rules:
//
// 1) Letter:
// ALetter, Hebrew_Letter, Numeric, ExtendNumLet, Extend, Format_FE, ZWJ.
// Rule: Never break between consecutive runes of this category.
//
// 2) Mid:
// MidLetter, MidNumLet, Single_Quote.
// (Cf. case-ignorable: MidLetter, MidNumLet, Single_Quote or cat is Mn,
// Me, Cf, Lm or Sk).
// Rule: Don't break between Letter and Mid, but break between two Mids.
//
// 3) Break:
// Any other category: NewLine, MidNum, CR, LF, Double_Quote, Katakana, and
// Other.
// These categories should always result in a break between two cased letters.
// Rule: Always break.
//
// Note 1: the Katakana and MidNum categories can, in esoteric cases, result in
// preventing a break between two cased letters. For now we will ignore this
// (e.g. [ALetter] [ExtendNumLet] [Katakana] [ExtendNumLet] [ALetter] and
// [ALetter] [Numeric] [MidNum] [Numeric] [ALetter].)
//
// Note 2: the rule for Mid is very approximate, but works in most cases. To
// improve, we could store the categories in the trie value and use a FA to
// manage breaks. See TODO comment above.
//
// Note 3: according to the spec, it is possible for the Extend category to
// introduce breaks between other categories grouped in Letter. However, this
// is undesirable for our purposes. ICU prevents breaks in such cases as well.
// isBreak returns whether this rune should introduce a break.
func (c info) isBreak() bool {
return c.cccVal() == cccBreak
}
// isLetter returns whether the rune is of break type ALetter, Hebrew_Letter,
// Numeric, ExtendNumLet, or Extend.
func (c info) isLetter() bool {
ccc := c.cccVal()
if ccc == cccZero {
return !c.isCaseIgnorable()
}
return ccc != cccBreak
}

816
vendor/golang.org/x/text/cases/map.go generated vendored
View file

@ -1,816 +0,0 @@
// Copyright 2014 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 cases
// This file contains the definitions of case mappings for all supported
// languages. The rules for the language-specific tailorings were taken and
// modified from the CLDR transform definitions in common/transforms.
import (
"strings"
"unicode"
"unicode/utf8"
"golang.org/x/text/internal"
"golang.org/x/text/language"
"golang.org/x/text/transform"
"golang.org/x/text/unicode/norm"
)
// A mapFunc takes a context set to the current rune and writes the mapped
// version to the same context. It may advance the context to the next rune. It
// returns whether a checkpoint is possible: whether the pDst bytes written to
// dst so far won't need changing as we see more source bytes.
type mapFunc func(*context) bool
// A spanFunc takes a context set to the current rune and returns whether this
// rune would be altered when written to the output. It may advance the context
// to the next rune. It returns whether a checkpoint is possible.
type spanFunc func(*context) bool
// maxIgnorable defines the maximum number of ignorables to consider for
// lookahead operations.
const maxIgnorable = 30
// supported lists the language tags for which we have tailorings.
const supported = "und af az el lt nl tr"
func init() {
tags := []language.Tag{}
for _, s := range strings.Split(supported, " ") {
tags = append(tags, language.MustParse(s))
}
matcher = internal.NewInheritanceMatcher(tags)
Supported = language.NewCoverage(tags)
}
var (
matcher *internal.InheritanceMatcher
Supported language.Coverage
// We keep the following lists separate, instead of having a single per-
// language struct, to give the compiler a chance to remove unused code.
// Some uppercase mappers are stateless, so we can precompute the
// Transformers and save a bit on runtime allocations.
upperFunc = []struct {
upper mapFunc
span spanFunc
}{
{nil, nil}, // und
{nil, nil}, // af
{aztrUpper(upper), isUpper}, // az
{elUpper, noSpan}, // el
{ltUpper(upper), noSpan}, // lt
{nil, nil}, // nl
{aztrUpper(upper), isUpper}, // tr
}
undUpper transform.SpanningTransformer = &undUpperCaser{}
undLower transform.SpanningTransformer = &undLowerCaser{}
undLowerIgnoreSigma transform.SpanningTransformer = &undLowerIgnoreSigmaCaser{}
lowerFunc = []mapFunc{
nil, // und
nil, // af
aztrLower, // az
nil, // el
ltLower, // lt
nil, // nl
aztrLower, // tr
}
titleInfos = []struct {
title mapFunc
lower mapFunc
titleSpan spanFunc
rewrite func(*context)
}{
{title, lower, isTitle, nil}, // und
{title, lower, isTitle, afnlRewrite}, // af
{aztrUpper(title), aztrLower, isTitle, nil}, // az
{title, lower, isTitle, nil}, // el
{ltUpper(title), ltLower, noSpan, nil}, // lt
{nlTitle, lower, nlTitleSpan, afnlRewrite}, // nl
{aztrUpper(title), aztrLower, isTitle, nil}, // tr
}
)
func makeUpper(t language.Tag, o options) transform.SpanningTransformer {
_, i, _ := matcher.Match(t)
f := upperFunc[i].upper
if f == nil {
return undUpper
}
return &simpleCaser{f: f, span: upperFunc[i].span}
}
func makeLower(t language.Tag, o options) transform.SpanningTransformer {
_, i, _ := matcher.Match(t)
f := lowerFunc[i]
if f == nil {
if o.ignoreFinalSigma {
return undLowerIgnoreSigma
}
return undLower
}
if o.ignoreFinalSigma {
return &simpleCaser{f: f, span: isLower}
}
return &lowerCaser{
first: f,
midWord: finalSigma(f),
}
}
func makeTitle(t language.Tag, o options) transform.SpanningTransformer {
_, i, _ := matcher.Match(t)
x := &titleInfos[i]
lower := x.lower
if o.noLower {
lower = (*context).copy
} else if !o.ignoreFinalSigma {
lower = finalSigma(lower)
}
return &titleCaser{
title: x.title,
lower: lower,
titleSpan: x.titleSpan,
rewrite: x.rewrite,
}
}
func noSpan(c *context) bool {
c.err = transform.ErrEndOfSpan
return false
}
// TODO: consider a similar special case for the fast majority lower case. This
// is a bit more involved so will require some more precise benchmarking to
// justify it.
type undUpperCaser struct{ transform.NopResetter }
// undUpperCaser implements the Transformer interface for doing an upper case
// mapping for the root locale (und). It eliminates the need for an allocation
// as it prevents escaping by not using function pointers.
func (t undUpperCaser) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
c := context{dst: dst, src: src, atEOF: atEOF}
for c.next() {
upper(&c)
c.checkpoint()
}
return c.ret()
}
func (t undUpperCaser) Span(src []byte, atEOF bool) (n int, err error) {
c := context{src: src, atEOF: atEOF}
for c.next() && isUpper(&c) {
c.checkpoint()
}
return c.retSpan()
}
// undLowerIgnoreSigmaCaser implements the Transformer interface for doing
// a lower case mapping for the root locale (und) ignoring final sigma
// handling. This casing algorithm is used in some performance-critical packages
// like secure/precis and x/net/http/idna, which warrants its special-casing.
type undLowerIgnoreSigmaCaser struct{ transform.NopResetter }
func (t undLowerIgnoreSigmaCaser) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
c := context{dst: dst, src: src, atEOF: atEOF}
for c.next() && lower(&c) {
c.checkpoint()
}
return c.ret()
}
// Span implements a generic lower-casing. This is possible as isLower works
// for all lowercasing variants. All lowercase variants only vary in how they
// transform a non-lowercase letter. They will never change an already lowercase
// letter. In addition, there is no state.
func (t undLowerIgnoreSigmaCaser) Span(src []byte, atEOF bool) (n int, err error) {
c := context{src: src, atEOF: atEOF}
for c.next() && isLower(&c) {
c.checkpoint()
}
return c.retSpan()
}
type simpleCaser struct {
context
f mapFunc
span spanFunc
}
// simpleCaser implements the Transformer interface for doing a case operation
// on a rune-by-rune basis.
func (t *simpleCaser) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
c := context{dst: dst, src: src, atEOF: atEOF}
for c.next() && t.f(&c) {
c.checkpoint()
}
return c.ret()
}
func (t *simpleCaser) Span(src []byte, atEOF bool) (n int, err error) {
c := context{src: src, atEOF: atEOF}
for c.next() && t.span(&c) {
c.checkpoint()
}
return c.retSpan()
}
// undLowerCaser implements the Transformer interface for doing a lower case
// mapping for the root locale (und) ignoring final sigma handling. This casing
// algorithm is used in some performance-critical packages like secure/precis
// and x/net/http/idna, which warrants its special-casing.
type undLowerCaser struct{ transform.NopResetter }
func (t undLowerCaser) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
c := context{dst: dst, src: src, atEOF: atEOF}
for isInterWord := true; c.next(); {
if isInterWord {
if c.info.isCased() {
if !lower(&c) {
break
}
isInterWord = false
} else if !c.copy() {
break
}
} else {
if c.info.isNotCasedAndNotCaseIgnorable() {
if !c.copy() {
break
}
isInterWord = true
} else if !c.hasPrefix("Σ") {
if !lower(&c) {
break
}
} else if !finalSigmaBody(&c) {
break
}
}
c.checkpoint()
}
return c.ret()
}
func (t undLowerCaser) Span(src []byte, atEOF bool) (n int, err error) {
c := context{src: src, atEOF: atEOF}
for c.next() && isLower(&c) {
c.checkpoint()
}
return c.retSpan()
}
// lowerCaser implements the Transformer interface. The default Unicode lower
// casing requires different treatment for the first and subsequent characters
// of a word, most notably to handle the Greek final Sigma.
type lowerCaser struct {
undLowerIgnoreSigmaCaser
context
first, midWord mapFunc
}
func (t *lowerCaser) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
t.context = context{dst: dst, src: src, atEOF: atEOF}
c := &t.context
for isInterWord := true; c.next(); {
if isInterWord {
if c.info.isCased() {
if !t.first(c) {
break
}
isInterWord = false
} else if !c.copy() {
break
}
} else {
if c.info.isNotCasedAndNotCaseIgnorable() {
if !c.copy() {
break
}
isInterWord = true
} else if !t.midWord(c) {
break
}
}
c.checkpoint()
}
return c.ret()
}
// titleCaser implements the Transformer interface. Title casing algorithms
// distinguish between the first letter of a word and subsequent letters of the
// same word. It uses state to avoid requiring a potentially infinite lookahead.
type titleCaser struct {
context
// rune mappings used by the actual casing algorithms.
title mapFunc
lower mapFunc
titleSpan spanFunc
rewrite func(*context)
}
// Transform implements the standard Unicode title case algorithm as defined in
// Chapter 3 of The Unicode Standard:
// toTitlecase(X): Find the word boundaries in X according to Unicode Standard
// Annex #29, "Unicode Text Segmentation." For each word boundary, find the
// first cased character F following the word boundary. If F exists, map F to
// Titlecase_Mapping(F); then map all characters C between F and the following
// word boundary to Lowercase_Mapping(C).
func (t *titleCaser) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
t.context = context{dst: dst, src: src, atEOF: atEOF, isMidWord: t.isMidWord}
c := &t.context
if !c.next() {
return c.ret()
}
for {
p := c.info
if t.rewrite != nil {
t.rewrite(c)
}
wasMid := p.isMid()
// Break out of this loop on failure to ensure we do not modify the
// state incorrectly.
if p.isCased() {
if !c.isMidWord {
if !t.title(c) {
break
}
c.isMidWord = true
} else if !t.lower(c) {
break
}
} else if !c.copy() {
break
} else if p.isBreak() {
c.isMidWord = false
}
// As we save the state of the transformer, it is safe to call
// checkpoint after any successful write.
if !(c.isMidWord && wasMid) {
c.checkpoint()
}
if !c.next() {
break
}
if wasMid && c.info.isMid() {
c.isMidWord = false
}
}
return c.ret()
}
func (t *titleCaser) Span(src []byte, atEOF bool) (n int, err error) {
t.context = context{src: src, atEOF: atEOF, isMidWord: t.isMidWord}
c := &t.context
if !c.next() {
return c.retSpan()
}
for {
p := c.info
if t.rewrite != nil {
t.rewrite(c)
}
wasMid := p.isMid()
// Break out of this loop on failure to ensure we do not modify the
// state incorrectly.
if p.isCased() {
if !c.isMidWord {
if !t.titleSpan(c) {
break
}
c.isMidWord = true
} else if !isLower(c) {
break
}
} else if p.isBreak() {
c.isMidWord = false
}
// As we save the state of the transformer, it is safe to call
// checkpoint after any successful write.
if !(c.isMidWord && wasMid) {
c.checkpoint()
}
if !c.next() {
break
}
if wasMid && c.info.isMid() {
c.isMidWord = false
}
}
return c.retSpan()
}
// finalSigma adds Greek final Sigma handing to another casing function. It
// determines whether a lowercased sigma should be σ or ς, by looking ahead for
// case-ignorables and a cased letters.
func finalSigma(f mapFunc) mapFunc {
return func(c *context) bool {
if !c.hasPrefix("Σ") {
return f(c)
}
return finalSigmaBody(c)
}
}
func finalSigmaBody(c *context) bool {
// Current rune must be ∑.
// ::NFD();
// # 03A3; 03C2; 03A3; 03A3; Final_Sigma; # GREEK CAPITAL LETTER SIGMA
// Σ } [:case-ignorable:]* [:cased:] → σ;
// [:cased:] [:case-ignorable:]* { Σ → ς;
// ::Any-Lower;
// ::NFC();
p := c.pDst
c.writeString("ς")
// TODO: we should do this here, but right now this will never have an
// effect as this is called when the prefix is Sigma, whereas Dutch and
// Afrikaans only test for an apostrophe.
//
// if t.rewrite != nil {
// t.rewrite(c)
// }
// We need to do one more iteration after maxIgnorable, as a cased
// letter is not an ignorable and may modify the result.
wasMid := false
for i := 0; i < maxIgnorable+1; i++ {
if !c.next() {
return false
}
if !c.info.isCaseIgnorable() {
// All Midword runes are also case ignorable, so we are
// guaranteed to have a letter or word break here. As we are
// unreading the run, there is no need to unset c.isMidWord;
// the title caser will handle this.
if c.info.isCased() {
// p+1 is guaranteed to be in bounds: if writing ς was
// successful, p+1 will contain the second byte of ς. If not,
// this function will have returned after c.next returned false.
c.dst[p+1]++ // ς → σ
}
c.unreadRune()
return true
}
// A case ignorable may also introduce a word break, so we may need
// to continue searching even after detecting a break.
isMid := c.info.isMid()
if (wasMid && isMid) || c.info.isBreak() {
c.isMidWord = false
}
wasMid = isMid
c.copy()
}
return true
}
// finalSigmaSpan would be the same as isLower.
// elUpper implements Greek upper casing, which entails removing a predefined
// set of non-blocked modifiers. Note that these accents should not be removed
// for title casing!
// Example: "Οδός" -> "ΟΔΟΣ".
func elUpper(c *context) bool {
// From CLDR:
// [:Greek:] [^[:ccc=Not_Reordered:][:ccc=Above:]]*? { [\u0313\u0314\u0301\u0300\u0306\u0342\u0308\u0304] → ;
// [:Greek:] [^[:ccc=Not_Reordered:][:ccc=Iota_Subscript:]]*? { \u0345 → ;
r, _ := utf8.DecodeRune(c.src[c.pSrc:])
oldPDst := c.pDst
if !upper(c) {
return false
}
if !unicode.Is(unicode.Greek, r) {
return true
}
i := 0
// Take the properties of the uppercased rune that is already written to the
// destination. This saves us the trouble of having to uppercase the
// decomposed rune again.
if b := norm.NFD.Properties(c.dst[oldPDst:]).Decomposition(); b != nil {
// Restore the destination position and process the decomposed rune.
r, sz := utf8.DecodeRune(b)
if r <= 0xFF { // See A.6.1
return true
}
c.pDst = oldPDst
// Insert the first rune and ignore the modifiers. See A.6.2.
c.writeBytes(b[:sz])
i = len(b[sz:]) / 2 // Greek modifiers are always of length 2.
}
for ; i < maxIgnorable && c.next(); i++ {
switch r, _ := utf8.DecodeRune(c.src[c.pSrc:]); r {
// Above and Iota Subscript
case 0x0300, // U+0300 COMBINING GRAVE ACCENT
0x0301, // U+0301 COMBINING ACUTE ACCENT
0x0304, // U+0304 COMBINING MACRON
0x0306, // U+0306 COMBINING BREVE
0x0308, // U+0308 COMBINING DIAERESIS
0x0313, // U+0313 COMBINING COMMA ABOVE
0x0314, // U+0314 COMBINING REVERSED COMMA ABOVE
0x0342, // U+0342 COMBINING GREEK PERISPOMENI
0x0345: // U+0345 COMBINING GREEK YPOGEGRAMMENI
// No-op. Gobble the modifier.
default:
switch v, _ := trie.lookup(c.src[c.pSrc:]); info(v).cccType() {
case cccZero:
c.unreadRune()
return true
// We don't need to test for IotaSubscript as the only rune that
// qualifies (U+0345) was already excluded in the switch statement
// above. See A.4.
case cccAbove:
return c.copy()
default:
// Some other modifier. We're still allowed to gobble Greek
// modifiers after this.
c.copy()
}
}
}
return i == maxIgnorable
}
// TODO: implement elUpperSpan (low-priority: complex and infrequent).
func ltLower(c *context) bool {
// From CLDR:
// # Introduce an explicit dot above when lowercasing capital I's and J's
// # whenever there are more accents above.
// # (of the accents used in Lithuanian: grave, acute, tilde above, and ogonek)
// # 0049; 0069 0307; 0049; 0049; lt More_Above; # LATIN CAPITAL LETTER I
// # 004A; 006A 0307; 004A; 004A; lt More_Above; # LATIN CAPITAL LETTER J
// # 012E; 012F 0307; 012E; 012E; lt More_Above; # LATIN CAPITAL LETTER I WITH OGONEK
// # 00CC; 0069 0307 0300; 00CC; 00CC; lt; # LATIN CAPITAL LETTER I WITH GRAVE
// # 00CD; 0069 0307 0301; 00CD; 00CD; lt; # LATIN CAPITAL LETTER I WITH ACUTE
// # 0128; 0069 0307 0303; 0128; 0128; lt; # LATIN CAPITAL LETTER I WITH TILDE
// ::NFD();
// I } [^[:ccc=Not_Reordered:][:ccc=Above:]]* [:ccc=Above:] → i \u0307;
// J } [^[:ccc=Not_Reordered:][:ccc=Above:]]* [:ccc=Above:] → j \u0307;
// I \u0328 (Į) } [^[:ccc=Not_Reordered:][:ccc=Above:]]* [:ccc=Above:] → i \u0328 \u0307;
// I \u0300 (Ì) → i \u0307 \u0300;
// I \u0301 (Í) → i \u0307 \u0301;
// I \u0303 (Ĩ) → i \u0307 \u0303;
// ::Any-Lower();
// ::NFC();
i := 0
if r := c.src[c.pSrc]; r < utf8.RuneSelf {
lower(c)
if r != 'I' && r != 'J' {
return true
}
} else {
p := norm.NFD.Properties(c.src[c.pSrc:])
if d := p.Decomposition(); len(d) >= 3 && (d[0] == 'I' || d[0] == 'J') {
// UTF-8 optimization: the decomposition will only have an above
// modifier if the last rune of the decomposition is in [U+300-U+311].
// In all other cases, a decomposition starting with I is always
// an I followed by modifiers that are not cased themselves. See A.2.
if d[1] == 0xCC && d[2] <= 0x91 { // A.2.4.
if !c.writeBytes(d[:1]) {
return false
}
c.dst[c.pDst-1] += 'a' - 'A' // lower
// Assumption: modifier never changes on lowercase. See A.1.
// Assumption: all modifiers added have CCC = Above. See A.2.3.
return c.writeString("\u0307") && c.writeBytes(d[1:])
}
// In all other cases the additional modifiers will have a CCC
// that is less than 230 (Above). We will insert the U+0307, if
// needed, after these modifiers so that a string in FCD form
// will remain so. See A.2.2.
lower(c)
i = 1
} else {
return lower(c)
}
}
for ; i < maxIgnorable && c.next(); i++ {
switch c.info.cccType() {
case cccZero:
c.unreadRune()
return true
case cccAbove:
return c.writeString("\u0307") && c.copy() // See A.1.
default:
c.copy() // See A.1.
}
}
return i == maxIgnorable
}
// ltLowerSpan would be the same as isLower.
func ltUpper(f mapFunc) mapFunc {
return func(c *context) bool {
// Unicode:
// 0307; 0307; ; ; lt After_Soft_Dotted; # COMBINING DOT ABOVE
//
// From CLDR:
// # Remove \u0307 following soft-dotteds (i, j, and the like), with possible
// # intervening non-230 marks.
// ::NFD();
// [:Soft_Dotted:] [^[:ccc=Not_Reordered:][:ccc=Above:]]* { \u0307 → ;
// ::Any-Upper();
// ::NFC();
// TODO: See A.5. A soft-dotted rune never has an exception. This would
// allow us to overload the exception bit and encode this property in
// info. Need to measure performance impact of this.
r, _ := utf8.DecodeRune(c.src[c.pSrc:])
oldPDst := c.pDst
if !f(c) {
return false
}
if !unicode.Is(unicode.Soft_Dotted, r) {
return true
}
// We don't need to do an NFD normalization, as a soft-dotted rune never
// contains U+0307. See A.3.
i := 0
for ; i < maxIgnorable && c.next(); i++ {
switch c.info.cccType() {
case cccZero:
c.unreadRune()
return true
case cccAbove:
if c.hasPrefix("\u0307") {
// We don't do a full NFC, but rather combine runes for
// some of the common cases. (Returning NFC or
// preserving normal form is neither a requirement nor
// a possibility anyway).
if !c.next() {
return false
}
if c.dst[oldPDst] == 'I' && c.pDst == oldPDst+1 && c.src[c.pSrc] == 0xcc {
s := ""
switch c.src[c.pSrc+1] {
case 0x80: // U+0300 COMBINING GRAVE ACCENT
s = "\u00cc" // U+00CC LATIN CAPITAL LETTER I WITH GRAVE
case 0x81: // U+0301 COMBINING ACUTE ACCENT
s = "\u00cd" // U+00CD LATIN CAPITAL LETTER I WITH ACUTE
case 0x83: // U+0303 COMBINING TILDE
s = "\u0128" // U+0128 LATIN CAPITAL LETTER I WITH TILDE
case 0x88: // U+0308 COMBINING DIAERESIS
s = "\u00cf" // U+00CF LATIN CAPITAL LETTER I WITH DIAERESIS
default:
}
if s != "" {
c.pDst = oldPDst
return c.writeString(s)
}
}
}
return c.copy()
default:
c.copy()
}
}
return i == maxIgnorable
}
}
// TODO: implement ltUpperSpan (low priority: complex and infrequent).
func aztrUpper(f mapFunc) mapFunc {
return func(c *context) bool {
// i→İ;
if c.src[c.pSrc] == 'i' {
return c.writeString("İ")
}
return f(c)
}
}
func aztrLower(c *context) (done bool) {
// From CLDR:
// # I and i-dotless; I-dot and i are case pairs in Turkish and Azeri
// # 0130; 0069; 0130; 0130; tr; # LATIN CAPITAL LETTER I WITH DOT ABOVE
// İ→i;
// # When lowercasing, remove dot_above in the sequence I + dot_above, which will turn into i.
// # This matches the behavior of the canonically equivalent I-dot_above
// # 0307; ; 0307; 0307; tr After_I; # COMBINING DOT ABOVE
// # When lowercasing, unless an I is before a dot_above, it turns into a dotless i.
// # 0049; 0131; 0049; 0049; tr Not_Before_Dot; # LATIN CAPITAL LETTER I
// I([^[:ccc=Not_Reordered:][:ccc=Above:]]*)\u0307 → i$1 ;
// I→ı ;
// ::Any-Lower();
if c.hasPrefix("\u0130") { // İ
return c.writeString("i")
}
if c.src[c.pSrc] != 'I' {
return lower(c)
}
// We ignore the lower-case I for now, but insert it later when we know
// which form we need.
start := c.pSrc + c.sz
i := 0
Loop:
// We check for up to n ignorables before \u0307. As \u0307 is an
// ignorable as well, n is maxIgnorable-1.
for ; i < maxIgnorable && c.next(); i++ {
switch c.info.cccType() {
case cccAbove:
if c.hasPrefix("\u0307") {
return c.writeString("i") && c.writeBytes(c.src[start:c.pSrc]) // ignore U+0307
}
done = true
break Loop
case cccZero:
c.unreadRune()
done = true
break Loop
default:
// We'll write this rune after we know which starter to use.
}
}
if i == maxIgnorable {
done = true
}
return c.writeString("ı") && c.writeBytes(c.src[start:c.pSrc+c.sz]) && done
}
// aztrLowerSpan would be the same as isLower.
func nlTitle(c *context) bool {
// From CLDR:
// # Special titlecasing for Dutch initial "ij".
// ::Any-Title();
// # Fix up Ij at the beginning of a "word" (per Any-Title, notUAX #29)
// [:^WB=ALetter:] [:WB=Extend:]* [[:WB=MidLetter:][:WB=MidNumLet:]]? { Ij } → IJ ;
if c.src[c.pSrc] != 'I' && c.src[c.pSrc] != 'i' {
return title(c)
}
if !c.writeString("I") || !c.next() {
return false
}
if c.src[c.pSrc] == 'j' || c.src[c.pSrc] == 'J' {
return c.writeString("J")
}
c.unreadRune()
return true
}
func nlTitleSpan(c *context) bool {
// From CLDR:
// # Special titlecasing for Dutch initial "ij".
// ::Any-Title();
// # Fix up Ij at the beginning of a "word" (per Any-Title, notUAX #29)
// [:^WB=ALetter:] [:WB=Extend:]* [[:WB=MidLetter:][:WB=MidNumLet:]]? { Ij } → IJ ;
if c.src[c.pSrc] != 'I' {
return isTitle(c)
}
if !c.next() || c.src[c.pSrc] == 'j' {
return false
}
if c.src[c.pSrc] != 'J' {
c.unreadRune()
}
return true
}
// Not part of CLDR, but see http://unicode.org/cldr/trac/ticket/7078.
func afnlRewrite(c *context) {
if c.hasPrefix("'") || c.hasPrefix("’") {
c.isMidWord = true
}
}

2211
vendor/golang.org/x/text/cases/tables.go generated vendored
View file

File diff suppressed because it is too large Load diff

215
vendor/golang.org/x/text/cases/trieval.go generated vendored
View file

@ -1,215 +0,0 @@
// Code generated by running "go generate" in golang.org/x/text. DO NOT EDIT.
package cases
// This file contains definitions for interpreting the trie value of the case
// trie generated by "go run gen*.go". It is shared by both the generator
// program and the resultant package. Sharing is achieved by the generator
// copying gen_trieval.go to trieval.go and changing what's above this comment.
// info holds case information for a single rune. It is the value returned
// by a trie lookup. Most mapping information can be stored in a single 16-bit
// value. If not, for example when a rune is mapped to multiple runes, the value
// stores some basic case data and an index into an array with additional data.
//
// The per-rune values have the following format:
//
// if (exception) {
// 15..5 unsigned exception index
// 4 unused
// } else {
// 15..8 XOR pattern or index to XOR pattern for case mapping
// Only 13..8 are used for XOR patterns.
// 7 inverseFold (fold to upper, not to lower)
// 6 index: interpret the XOR pattern as an index
// or isMid if case mode is cIgnorableUncased.
// 5..4 CCC: zero (normal or break), above or other
// }
// 3 exception: interpret this value as an exception index
// (TODO: is this bit necessary? Probably implied from case mode.)
// 2..0 case mode
//
// For the non-exceptional cases, a rune must be either uncased, lowercase or
// uppercase. If the rune is cased, the XOR pattern maps either a lowercase
// rune to uppercase or an uppercase rune to lowercase (applied to the 10
// least-significant bits of the rune).
//
// See the definitions below for a more detailed description of the various
// bits.
type info uint16
const (
casedMask = 0x0003
fullCasedMask = 0x0007
ignorableMask = 0x0006
ignorableValue = 0x0004
inverseFoldBit = 1 << 7
isMidBit = 1 << 6
exceptionBit = 1 << 3
exceptionShift = 5
numExceptionBits = 11
xorIndexBit = 1 << 6
xorShift = 8
// There is no mapping if all xor bits and the exception bit are zero.
hasMappingMask = 0xff80 | exceptionBit
)
// The case mode bits encodes the case type of a rune. This includes uncased,
// title, upper and lower case and case ignorable. (For a definition of these
// terms see Chapter 3 of The Unicode Standard Core Specification.) In some rare
// cases, a rune can be both cased and case-ignorable. This is encoded by
// cIgnorableCased. A rune of this type is always lower case. Some runes are
// cased while not having a mapping.
//
// A common pattern for scripts in the Unicode standard is for upper and lower
// case runes to alternate for increasing rune values (e.g. the accented Latin
// ranges starting from U+0100 and U+1E00 among others and some Cyrillic
// characters). We use this property by defining a cXORCase mode, where the case
// mode (always upper or lower case) is derived from the rune value. As the XOR
// pattern for case mappings is often identical for successive runes, using
// cXORCase can result in large series of identical trie values. This, in turn,
// allows us to better compress the trie blocks.
const (
cUncased info = iota // 000
cTitle // 001
cLower // 010
cUpper // 011
cIgnorableUncased // 100
cIgnorableCased // 101 // lower case if mappings exist
cXORCase // 11x // case is cLower | ((rune&1) ^ x)
maxCaseMode = cUpper
)
func (c info) isCased() bool {
return c&casedMask != 0
}
func (c info) isCaseIgnorable() bool {
return c&ignorableMask == ignorableValue
}
func (c info) isNotCasedAndNotCaseIgnorable() bool {
return c&fullCasedMask == 0
}
func (c info) isCaseIgnorableAndNotCased() bool {
return c&fullCasedMask == cIgnorableUncased
}
func (c info) isMid() bool {
return c&(fullCasedMask|isMidBit) == isMidBit|cIgnorableUncased
}
// The case mapping implementation will need to know about various Canonical
// Combining Class (CCC) values. We encode two of these in the trie value:
// cccZero (0) and cccAbove (230). If the value is cccOther, it means that
// CCC(r) > 0, but not 230. A value of cccBreak means that CCC(r) == 0 and that
// the rune also has the break category Break (see below).
const (
cccBreak info = iota << 4
cccZero
cccAbove
cccOther
cccMask = cccBreak | cccZero | cccAbove | cccOther
)
const (
starter = 0
above = 230
iotaSubscript = 240
)
// The exceptions slice holds data that does not fit in a normal info entry.
// The entry is pointed to by the exception index in an entry. It has the
// following format:
//
// Header
// byte 0:
// 7..6 unused
// 5..4 CCC type (same bits as entry)
// 3 unused
// 2..0 length of fold
//
// byte 1:
// 7..6 unused
// 5..3 length of 1st mapping of case type
// 2..0 length of 2nd mapping of case type
//
// case 1st 2nd
// lower -> upper, title
// upper -> lower, title
// title -> lower, upper
//
// Lengths with the value 0x7 indicate no value and implies no change.
// A length of 0 indicates a mapping to zero-length string.
//
// Body bytes:
// case folding bytes
// lowercase mapping bytes
// uppercase mapping bytes
// titlecase mapping bytes
// closure mapping bytes (for NFKC_Casefold). (TODO)
//
// Fallbacks:
// missing fold -> lower
// missing title -> upper
// all missing -> original rune
//
// exceptions starts with a dummy byte to enforce that there is no zero index
// value.
const (
lengthMask = 0x07
lengthBits = 3
noChange = 0
)
// References to generated trie.
var trie = newCaseTrie(0)
var sparse = sparseBlocks{
values: sparseValues[:],
offsets: sparseOffsets[:],
}
// Sparse block lookup code.
// valueRange is an entry in a sparse block.
type valueRange struct {
value uint16
lo, hi byte
}
type sparseBlocks struct {
values []valueRange
offsets []uint16
}
// lookup returns the value from values block n for byte b using binary search.
func (s *sparseBlocks) lookup(n uint32, b byte) uint16 {
lo := s.offsets[n]
hi := s.offsets[n+1]
for lo < hi {
m := lo + (hi-lo)/2
r := s.values[m]
if r.lo <= b && b <= r.hi {
return r.value
}
if b < r.lo {
hi = m
} else {
lo = m + 1
}
}
return 0
}
// lastRuneForTesting is the last rune used for testing. Everything after this
// is boring.
const lastRuneForTesting = rune(0x1FFFF)

52
vendor/golang.org/x/text/internal/gen.go generated vendored
View file

@ -1,52 +0,0 @@
// 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
import (
"log"
"golang.org/x/text/internal/gen"
"golang.org/x/text/language"
"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("tables.go", "internal")
// Create parents table.
parents := make([]uint16, language.NumCompactTags)
for _, loc := range data.Locales() {
tag := language.MustParse(loc)
index, ok := language.CompactIndex(tag)
if !ok {
continue
}
parentIndex := 0 // und
for p := tag.Parent(); p != language.Und; p = p.Parent() {
if x, ok := language.CompactIndex(p); ok {
parentIndex = x
break
}
}
parents[index] = uint16(parentIndex)
}
w.WriteComment(`
Parent maps a compact index of a tag to the compact index of the parent of
this tag.`)
w.WriteVar("Parent", parents)
}

51
vendor/golang.org/x/text/internal/internal.go generated vendored
View file

@ -1,51 +0,0 @@
// 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.
//go:generate go run gen.go
// Package internal contains non-exported functionality that are used by
// packages in the text repository.
package internal // import "golang.org/x/text/internal"
import (
"sort"
"golang.org/x/text/language"
)
// SortTags sorts tags in place.
func SortTags(tags []language.Tag) {
sort.Sort(sorter(tags))
}
type sorter []language.Tag
func (s sorter) Len() int {
return len(s)
}
func (s sorter) Swap(i, j int) {
s[i], s[j] = s[j], s[i]
}
func (s sorter) Less(i, j int) bool {
return s[i].String() < s[j].String()
}
// UniqueTags sorts and filters duplicate tags in place and returns a slice with
// only unique tags.
func UniqueTags(tags []language.Tag) []language.Tag {
if len(tags) <= 1 {
return tags
}
SortTags(tags)
k := 0
for i := 1; i < len(tags); i++ {
if tags[k].String() < tags[i].String() {
k++
tags[k] = tags[i]
}
}
return tags[:k+1]
}

67
vendor/golang.org/x/text/internal/match.go generated vendored
View file

@ -1,67 +0,0 @@
// 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.
package internal
// This file contains matchers that implement CLDR inheritance.
//
// See http://unicode.org/reports/tr35/#Locale_Inheritance.
//
// Some of the inheritance described in this document is already handled by
// the cldr package.
import (
"golang.org/x/text/language"
)
// TODO: consider if (some of the) matching algorithm needs to be public after
// getting some feel about what is generic and what is specific.
// NewInheritanceMatcher returns a matcher that matches based on the inheritance
// chain.
//
// The matcher uses canonicalization and the parent relationship to find a
// match. The resulting match will always be either Und or a language with the
// same language and script as the requested language. It will not match
// languages for which there is understood to be mutual or one-directional
// intelligibility.
//
// A Match will indicate an Exact match if the language matches after
// canonicalization and High if the matched tag is a parent.
func NewInheritanceMatcher(t []language.Tag) *InheritanceMatcher {
tags := &InheritanceMatcher{make(map[language.Tag]int)}
for i, tag := range t {
ct, err := language.All.Canonicalize(tag)
if err != nil {
ct = tag
}
tags.index[ct] = i
}
return tags
}
type InheritanceMatcher struct {
index map[language.Tag]int
}
func (m InheritanceMatcher) Match(want ...language.Tag) (language.Tag, int, language.Confidence) {
for _, t := range want {
ct, err := language.All.Canonicalize(t)
if err != nil {
ct = t
}
conf := language.Exact
for {
if index, ok := m.index[ct]; ok {
return ct, index, conf
}
if ct == language.Und {
break
}
ct = ct.Parent()
conf = language.High
}
}
return language.Und, 0, language.No
}

116
vendor/golang.org/x/text/internal/tables.go generated vendored
View file

@ -1,116 +0,0 @@
// Code generated by running "go generate" in golang.org/x/text. DO NOT EDIT.
package internal
// Parent maps a compact index of a tag to the compact index of the parent of
// this tag.
var Parent = []uint16{ // 752 elements
// Entry 0 - 3F
0x0000, 0x0053, 0x00e5, 0x0000, 0x0003, 0x0003, 0x0000, 0x0006,
0x0000, 0x0008, 0x0000, 0x000a, 0x0000, 0x000c, 0x000c, 0x000c,
0x000c, 0x000c, 0x000c, 0x000c, 0x000c, 0x000c, 0x000c, 0x000c,
0x000c, 0x000c, 0x000c, 0x000c, 0x000c, 0x000c, 0x000c, 0x000c,
0x000c, 0x000c, 0x000c, 0x000c, 0x000c, 0x000c, 0x000c, 0x000c,
0x000c, 0x0000, 0x0000, 0x002a, 0x0000, 0x002c, 0x0000, 0x002e,
0x0000, 0x0000, 0x0031, 0x0030, 0x0030, 0x0000, 0x0035, 0x0000,
0x0037, 0x0000, 0x0039, 0x0000, 0x003b, 0x0000, 0x003d, 0x0000,
// Entry 40 - 7F
0x0000, 0x0040, 0x0000, 0x0042, 0x0042, 0x0000, 0x0045, 0x0045,
0x0000, 0x0048, 0x0000, 0x004a, 0x0000, 0x0000, 0x004d, 0x004c,
0x004c, 0x0000, 0x0051, 0x0051, 0x0051, 0x0051, 0x0000, 0x0056,
0x0000, 0x0058, 0x0000, 0x005a, 0x0000, 0x005c, 0x005c, 0x0000,
0x005f, 0x0000, 0x0061, 0x0000, 0x0063, 0x0000, 0x0065, 0x0065,
0x0000, 0x0068, 0x0000, 0x006a, 0x006a, 0x006a, 0x006a, 0x006a,
0x006a, 0x006a, 0x0000, 0x0072, 0x0000, 0x0074, 0x0000, 0x0076,
0x0000, 0x0000, 0x0079, 0x0000, 0x007b, 0x0000, 0x007d, 0x0000,
// Entry 80 - BF
0x007f, 0x007f, 0x0000, 0x0082, 0x0082, 0x0000, 0x0085, 0x0086,
0x0086, 0x0086, 0x0085, 0x0087, 0x0086, 0x0086, 0x0086, 0x0085,
0x0086, 0x0086, 0x0086, 0x0086, 0x0086, 0x0086, 0x0087, 0x0086,
0x0086, 0x0086, 0x0086, 0x0087, 0x0086, 0x0087, 0x0086, 0x0086,
0x0087, 0x0086, 0x0086, 0x0086, 0x0086, 0x0086, 0x0086, 0x0086,
0x0086, 0x0086, 0x0085, 0x0086, 0x0086, 0x0086, 0x0086, 0x0086,
0x0086, 0x0086, 0x0086, 0x0086, 0x0086, 0x0086, 0x0086, 0x0086,
0x0086, 0x0086, 0x0086, 0x0086, 0x0085, 0x0086, 0x0085, 0x0086,
// Entry C0 - FF
0x0086, 0x0086, 0x0086, 0x0086, 0x0086, 0x0086, 0x0086, 0x0087,
0x0086, 0x0086, 0x0086, 0x0086, 0x0086, 0x0086, 0x0086, 0x0085,
0x0086, 0x0086, 0x0086, 0x0086, 0x0086, 0x0087, 0x0086, 0x0086,
0x0087, 0x0086, 0x0086, 0x0086, 0x0086, 0x0086, 0x0086, 0x0086,
0x0086, 0x0086, 0x0086, 0x0086, 0x0085, 0x0085, 0x0086, 0x0086,
0x0085, 0x0086, 0x0086, 0x0086, 0x0086, 0x0086, 0x0000, 0x00ee,
0x0000, 0x00f0, 0x00f1, 0x00f1, 0x00f1, 0x00f1, 0x00f1, 0x00f1,
0x00f1, 0x00f1, 0x00f0, 0x00f1, 0x00f0, 0x00f0, 0x00f1, 0x00f1,
// Entry 100 - 13F
0x00f0, 0x00f1, 0x00f1, 0x00f1, 0x00f1, 0x00f0, 0x00f1, 0x00f1,
0x00f1, 0x00f1, 0x00f1, 0x00f1, 0x0000, 0x010c, 0x0000, 0x010e,
0x0000, 0x0110, 0x0000, 0x0112, 0x0112, 0x0000, 0x0115, 0x0115,
0x0115, 0x0115, 0x0000, 0x011a, 0x0000, 0x011c, 0x0000, 0x011e,
0x011e, 0x0000, 0x0121, 0x0121, 0x0121, 0x0121, 0x0121, 0x0121,
0x0121, 0x0121, 0x0121, 0x0121, 0x0121, 0x0121, 0x0121, 0x0121,
0x0121, 0x0121, 0x0121, 0x0121, 0x0121, 0x0121, 0x0121, 0x0121,
0x0121, 0x0121, 0x0121, 0x0121, 0x0121, 0x0121, 0x0121, 0x0121,
// Entry 140 - 17F
0x0121, 0x0121, 0x0121, 0x0121, 0x0121, 0x0121, 0x0121, 0x0121,
0x0121, 0x0121, 0x0121, 0x0121, 0x0121, 0x0121, 0x0121, 0x0121,
0x0000, 0x0150, 0x0000, 0x0152, 0x0000, 0x0154, 0x0000, 0x0156,
0x0000, 0x0158, 0x0000, 0x015a, 0x015a, 0x015a, 0x0000, 0x015e,
0x0000, 0x0000, 0x0161, 0x0000, 0x0163, 0x0000, 0x0165, 0x0165,
0x0165, 0x0000, 0x0169, 0x0000, 0x016b, 0x0000, 0x016d, 0x0000,
0x016f, 0x016f, 0x0000, 0x0172, 0x0000, 0x0174, 0x0000, 0x0176,
0x0000, 0x0178, 0x0000, 0x017a, 0x0000, 0x017c, 0x0000, 0x017e,
// Entry 180 - 1BF
0x0000, 0x0180, 0x0180, 0x0180, 0x0000, 0x0000, 0x0185, 0x0000,
0x0000, 0x0188, 0x0000, 0x018a, 0x0000, 0x0000, 0x018d, 0x0000,
0x018f, 0x0000, 0x0000, 0x0192, 0x0000, 0x0000, 0x0195, 0x0000,
0x0197, 0x0000, 0x0199, 0x0000, 0x019b, 0x0000, 0x019d, 0x0000,
0x019f, 0x0000, 0x01a1, 0x0000, 0x01a3, 0x0000, 0x01a5, 0x0000,
0x01a7, 0x0000, 0x01a9, 0x01a9, 0x0000, 0x01ac, 0x0000, 0x01ae,
0x0000, 0x01b0, 0x0000, 0x01b2, 0x0000, 0x01b4, 0x0000, 0x0000,
0x01b7, 0x0000, 0x01b9, 0x0000, 0x01bb, 0x0000, 0x01bd, 0x0000,
// Entry 1C0 - 1FF
0x01bf, 0x0000, 0x01c1, 0x0000, 0x01c3, 0x01c3, 0x01c3, 0x01c3,
0x0000, 0x01c8, 0x0000, 0x01ca, 0x01ca, 0x0000, 0x01cd, 0x0000,
0x01cf, 0x0000, 0x01d1, 0x0000, 0x01d3, 0x0000, 0x01d5, 0x0000,
0x01d7, 0x01d7, 0x0000, 0x01da, 0x0000, 0x01dc, 0x0000, 0x01de,
0x0000, 0x01e0, 0x0000, 0x01e2, 0x0000, 0x01e4, 0x0000, 0x01e6,
0x0000, 0x01e8, 0x0000, 0x01ea, 0x0000, 0x01ec, 0x01ec, 0x01ec,
0x0000, 0x01f0, 0x0000, 0x01f2, 0x0000, 0x01f4, 0x0000, 0x01f6,
0x0000, 0x0000, 0x01f9, 0x0000, 0x01fb, 0x01fb, 0x0000, 0x01fe,
// Entry 200 - 23F
0x0000, 0x0200, 0x0200, 0x0000, 0x0203, 0x0203, 0x0000, 0x0206,
0x0206, 0x0206, 0x0206, 0x0206, 0x0206, 0x0206, 0x0000, 0x020e,
0x0000, 0x0210, 0x0000, 0x0212, 0x0000, 0x0000, 0x0000, 0x0000,
0x0000, 0x0218, 0x0000, 0x0000, 0x021b, 0x0000, 0x021d, 0x021d,
0x0000, 0x0220, 0x0000, 0x0222, 0x0222, 0x0000, 0x0000, 0x0226,
0x0225, 0x0225, 0x0000, 0x0000, 0x022b, 0x0000, 0x022d, 0x0000,
0x022f, 0x0000, 0x023b, 0x0231, 0x023b, 0x023b, 0x023b, 0x023b,
0x023b, 0x023b, 0x023b, 0x0231, 0x023b, 0x023b, 0x0000, 0x023e,
// Entry 240 - 27F
0x023e, 0x023e, 0x0000, 0x0242, 0x0000, 0x0244, 0x0000, 0x0246,
0x0246, 0x0000, 0x0249, 0x0000, 0x024b, 0x024b, 0x024b, 0x024b,
0x024b, 0x024b, 0x0000, 0x0252, 0x0000, 0x0254, 0x0000, 0x0256,
0x0000, 0x0258, 0x0000, 0x025a, 0x0000, 0x0000, 0x025d, 0x025d,
0x025d, 0x0000, 0x0261, 0x0000, 0x0263, 0x0000, 0x0265, 0x0000,
0x0000, 0x0268, 0x0267, 0x0267, 0x0000, 0x026c, 0x0000, 0x026e,
0x0000, 0x0270, 0x0000, 0x0000, 0x0000, 0x0000, 0x0275, 0x0000,
0x0000, 0x0278, 0x0000, 0x027a, 0x027a, 0x027a, 0x027a, 0x0000,
// Entry 280 - 2BF
0x027f, 0x027f, 0x027f, 0x0000, 0x0283, 0x0283, 0x0283, 0x0283,
0x0283, 0x0000, 0x0289, 0x0289, 0x0289, 0x0289, 0x0000, 0x0000,
0x0000, 0x0000, 0x0291, 0x0291, 0x0291, 0x0000, 0x0295, 0x0295,
0x0295, 0x0295, 0x0000, 0x0000, 0x029b, 0x029b, 0x029b, 0x029b,
0x0000, 0x02a0, 0x0000, 0x02a2, 0x02a2, 0x0000, 0x02a5, 0x0000,
0x02a7, 0x02a7, 0x0000, 0x0000, 0x02ab, 0x0000, 0x0000, 0x02ae,
0x0000, 0x02b0, 0x02b0, 0x0000, 0x0000, 0x02b4, 0x0000, 0x02b6,
0x0000, 0x02b8, 0x0000, 0x02ba, 0x0000, 0x02bc, 0x02bc, 0x0000,
// Entry 2C0 - 2FF
0x0000, 0x02c0, 0x0000, 0x02c2, 0x02bf, 0x02bf, 0x0000, 0x0000,
0x02c7, 0x02c6, 0x02c6, 0x0000, 0x0000, 0x02cc, 0x0000, 0x02ce,
0x0000, 0x02d0, 0x0000, 0x0000, 0x02d3, 0x0000, 0x0000, 0x0000,
0x02d7, 0x0000, 0x02d9, 0x0000, 0x02db, 0x0000, 0x02dd, 0x02dd,
0x0000, 0x02e0, 0x0000, 0x02e2, 0x0000, 0x02e4, 0x02e4, 0x02e4,
0x02e4, 0x02e4, 0x0000, 0x02ea, 0x02eb, 0x02ea, 0x0000, 0x02ee,
} // Size: 1528 bytes
// Total table size 1528 bytes (1KiB); checksum: B99CF952

100
vendor/golang.org/x/text/internal/tag/tag.go generated vendored
View file

@ -1,100 +0,0 @@
// 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.
// Package tag contains functionality handling tags and related data.
package tag // import "golang.org/x/text/internal/tag"
import "sort"
// An Index converts tags to a compact numeric value.
//
// All elements are of size 4. Tags may be up to 4 bytes long. Excess bytes can
// be used to store additional information about the tag.
type Index string
// Elem returns the element data at the given index.
func (s Index) Elem(x int) string {
return string(s[x*4 : x*4+4])
}
// Index reports the index of the given key or -1 if it could not be found.
// Only the first len(key) bytes from the start of the 4-byte entries will be
// considered for the search and the first match in Index will be returned.
func (s Index) Index(key []byte) int {
n := len(key)
// search the index of the first entry with an equal or higher value than
// key in s.
index := sort.Search(len(s)/4, func(i int) bool {
return cmp(s[i*4:i*4+n], key) != -1
})
i := index * 4
if cmp(s[i:i+len(key)], key) != 0 {
return -1
}
return index
}
// Next finds the next occurrence of key after index x, which must have been
// obtained from a call to Index using the same key. It returns x+1 or -1.
func (s Index) Next(key []byte, x int) int {
if x++; x*4 < len(s) && cmp(s[x*4:x*4+len(key)], key) == 0 {
return x
}
return -1
}
// cmp returns an integer comparing a and b lexicographically.
func cmp(a Index, b []byte) int {
n := len(a)
if len(b) < n {
n = len(b)
}
for i, c := range b[:n] {
switch {
case a[i] > c:
return 1
case a[i] < c:
return -1
}
}
switch {
case len(a) < len(b):
return -1
case len(a) > len(b):
return 1
}
return 0
}
// Compare returns an integer comparing a and b lexicographically.
func Compare(a string, b []byte) int {
return cmp(Index(a), b)
}
// FixCase reformats b to the same pattern of cases as form.
// If returns false if string b is malformed.
func FixCase(form string, b []byte) bool {
if len(form) != len(b) {
return false
}
for i, c := range b {
if form[i] <= 'Z' {
if c >= 'a' {
c -= 'z' - 'Z'
}
if c < 'A' || 'Z' < c {
return false
}
} else {
if c <= 'Z' {
c += 'z' - 'Z'
}
if c < 'a' || 'z' < c {
return false
}
}
b[i] = c
}
return true
}

16
vendor/golang.org/x/text/language/common.go generated vendored
View file

@ -1,16 +0,0 @@
// Code generated by running "go generate" in golang.org/x/text. DO NOT EDIT.
package language
// This file contains code common to the maketables.go and the package code.
// langAliasType is the type of an alias in langAliasMap.
type langAliasType int8
const (
langDeprecated langAliasType = iota
langMacro
langLegacy
langAliasTypeUnknown langAliasType = -1
)

197
vendor/golang.org/x/text/language/coverage.go generated vendored
View file

@ -1,197 +0,0 @@
// Copyright 2014 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 language
import (
"fmt"
"sort"
)
// The Coverage interface is used to define the level of coverage of an
// internationalization service. Note that not all types are supported by all
// services. As lists may be generated on the fly, it is recommended that users
// of a Coverage cache the results.
type Coverage interface {
// Tags returns the list of supported tags.
Tags() []Tag
// BaseLanguages returns the list of supported base languages.
BaseLanguages() []Base
// Scripts returns the list of supported scripts.
Scripts() []Script
// Regions returns the list of supported regions.
Regions() []Region
}
var (
// Supported defines a Coverage that lists all supported subtags. Tags
// always returns nil.
Supported Coverage = allSubtags{}
)
// TODO:
// - Support Variants, numbering systems.
// - CLDR coverage levels.
// - Set of common tags defined in this package.
type allSubtags struct{}
// Regions returns the list of supported regions. As all regions are in a
// consecutive range, it simply returns a slice of numbers in increasing order.
// The "undefined" region is not returned.
func (s allSubtags) Regions() []Region {
reg := make([]Region, numRegions)
for i := range reg {
reg[i] = Region{regionID(i + 1)}
}
return reg
}
// Scripts returns the list of supported scripts. As all scripts are in a
// consecutive range, it simply returns a slice of numbers in increasing order.
// The "undefined" script is not returned.
func (s allSubtags) Scripts() []Script {
scr := make([]Script, numScripts)
for i := range scr {
scr[i] = Script{scriptID(i + 1)}
}
return scr
}
// BaseLanguages returns the list of all supported base languages. It generates
// the list by traversing the internal structures.
func (s allSubtags) BaseLanguages() []Base {
base := make([]Base, 0, numLanguages)
for i := 0; i < langNoIndexOffset; i++ {
// We included "und" already for the value 0.
if i != nonCanonicalUnd {
base = append(base, Base{langID(i)})
}
}
i := langNoIndexOffset
for _, v := range langNoIndex {
for k := 0; k < 8; k++ {
if v&1 == 1 {
base = append(base, Base{langID(i)})
}
v >>= 1
i++
}
}
return base
}
// Tags always returns nil.
func (s allSubtags) Tags() []Tag {
return nil
}
// coverage is used used by NewCoverage which is used as a convenient way for
// creating Coverage implementations for partially defined data. Very often a
// package will only need to define a subset of slices. coverage provides a
// convenient way to do this. Moreover, packages using NewCoverage, instead of
// their own implementation, will not break if later new slice types are added.
type coverage struct {
tags func() []Tag
bases func() []Base
scripts func() []Script
regions func() []Region
}
func (s *coverage) Tags() []Tag {
if s.tags == nil {
return nil
}
return s.tags()
}
// bases implements sort.Interface and is used to sort base languages.
type bases []Base
func (b bases) Len() int {
return len(b)
}
func (b bases) Swap(i, j int) {
b[i], b[j] = b[j], b[i]
}
func (b bases) Less(i, j int) bool {
return b[i].langID < b[j].langID
}
// BaseLanguages returns the result from calling s.bases if it is specified or
// otherwise derives the set of supported base languages from tags.
func (s *coverage) BaseLanguages() []Base {
if s.bases == nil {
tags := s.Tags()
if len(tags) == 0 {
return nil
}
a := make([]Base, len(tags))
for i, t := range tags {
a[i] = Base{langID(t.lang)}
}
sort.Sort(bases(a))
k := 0
for i := 1; i < len(a); i++ {
if a[k] != a[i] {
k++
a[k] = a[i]
}
}
return a[:k+1]
}
return s.bases()
}
func (s *coverage) Scripts() []Script {
if s.scripts == nil {
return nil
}
return s.scripts()
}
func (s *coverage) Regions() []Region {
if s.regions == nil {
return nil
}
return s.regions()
}
// NewCoverage returns a Coverage for the given lists. It is typically used by
// packages providing internationalization services to define their level of
// coverage. A list may be of type []T or func() []T, where T is either Tag,
// Base, Script or Region. The returned Coverage derives the value for Bases
// from Tags if no func or slice for []Base is specified. For other unspecified
// types the returned Coverage will return nil for the respective methods.
func NewCoverage(list ...interface{}) Coverage {
s := &coverage{}
for _, x := range list {
switch v := x.(type) {
case func() []Base:
s.bases = v
case func() []Script:
s.scripts = v
case func() []Region:
s.regions = v
case func() []Tag:
s.tags = v
case []Base:
s.bases = func() []Base { return v }
case []Script:
s.scripts = func() []Script { return v }
case []Region:
s.regions = func() []Region { return v }
case []Tag:
s.tags = func() []Tag { return v }
default:
panic(fmt.Sprintf("language: unsupported set type %T", v))
}
}
return s
}

20
vendor/golang.org/x/text/language/gen_common.go generated vendored
View file

@ -1,20 +0,0 @@
// Copyright 2014 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 contains code common to the maketables.go and the package code.
// langAliasType is the type of an alias in langAliasMap.
type langAliasType int8
const (
langDeprecated langAliasType = iota
langMacro
langLegacy
langAliasTypeUnknown langAliasType = -1
)

162
vendor/golang.org/x/text/language/gen_index.go generated vendored
View file

@ -1,162 +0,0 @@
// 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 (
"bytes"
"flag"
"fmt"
"io/ioutil"
"log"
"reflect"
"sort"
"strings"
"golang.org/x/text/internal/gen"
"golang.org/x/text/language"
"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.")
draft = flag.String("draft",
"contributed",
`Minimal draft requirements (approved, contributed, provisional, unconfirmed).`)
)
func main() {
gen.Init()
// Read the CLDR zip file.
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 func() {
buf := &bytes.Buffer{}
if _, err = w.WriteGo(buf, "language"); err != nil {
log.Fatalf("Error formatting file index.go: %v", err)
}
// Since we're generating a table for our own package we need to rewrite
// doing the equivalent of go fmt -r 'language.b -> b'. Using
// bytes.Replace will do.
out := bytes.Replace(buf.Bytes(), []byte("language."), nil, -1)
if err := ioutil.WriteFile("index.go", out, 0600); err != nil {
log.Fatalf("Could not create file index.go: %v", err)
}
}()
m := map[language.Tag]bool{}
for _, lang := range 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
// }
}
// Include locales for plural rules, which uses a different structure.
for _, plurals := range data.Supplemental().Plurals {
for _, rules := range plurals.PluralRules {
for _, lang := range strings.Split(rules.Locales, " ") {
m[language.Make(lang)] = true
}
}
}
var core, special []language.Tag
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 {
core = append(core, t)
} else {
special = append(special, t)
}
}
w.WriteComment(`
NumCompactTags is the number of common tags. The maximum tag is
NumCompactTags-1.`)
w.WriteConst("NumCompactTags", len(core)+len(special))
sort.Sort(byAlpha(special))
w.WriteVar("specialTags", special)
// TODO: order by frequency?
sort.Sort(byAlpha(core))
// Size computations are just an estimate.
w.Size += int(reflect.TypeOf(map[uint32]uint16{}).Size())
w.Size += len(core) * 6 // size of uint32 and uint16
fmt.Fprintln(w)
fmt.Fprintln(w, "var coreTags = map[uint32]uint16{")
fmt.Fprintln(w, "0x0: 0, // und")
i := len(special) + 1 // Und and special tags already written.
for _, t := range core {
if t == language.Und {
continue
}
fmt.Fprint(w.Hash, t, i)
b, s, r := t.Raw()
fmt.Fprintf(w, "0x%s%s%s: %d, // %s\n",
getIndex(b, 3), // 3 is enough as it is guaranteed to be a compact number
getIndex(s, 2),
getIndex(r, 3),
i, t)
i++
}
fmt.Fprintln(w, "}")
}
// getIndex prints the subtag type and extracts its index of size nibble.
// If the index is less than n nibbles, the result is prefixed with 0s.
func getIndex(x interface{}, n int) string {
s := fmt.Sprintf("%#v", x) // s is of form Type{typeID: 0x00}
s = s[strings.Index(s, "0x")+2 : len(s)-1]
return strings.Repeat("0", n-len(s)) + s
}
type byAlpha []language.Tag
func (a byAlpha) Len() int { return len(a) }
func (a byAlpha) Swap(i, j int) { a[i], a[j] = a[j], a[i] }
func (a byAlpha) Less(i, j int) bool { return a[i].String() < a[j].String() }

38
vendor/golang.org/x/text/language/go1_1.go generated vendored
View file

@ -1,38 +0,0 @@
// 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 !go1.2
package language
import "sort"
func sortStable(s sort.Interface) {
ss := stableSort{
s: s,
pos: make([]int, s.Len()),
}
for i := range ss.pos {
ss.pos[i] = i
}
sort.Sort(&ss)
}
type stableSort struct {
s sort.Interface
pos []int
}
func (s *stableSort) Len() int {
return len(s.pos)
}
func (s *stableSort) Less(i, j int) bool {
return s.s.Less(i, j) || !s.s.Less(j, i) && s.pos[i] < s.pos[j]
}
func (s *stableSort) Swap(i, j int) {
s.s.Swap(i, j)
s.pos[i], s.pos[j] = s.pos[j], s.pos[i]
}

11
vendor/golang.org/x/text/language/go1_2.go generated vendored
View file

@ -1,11 +0,0 @@
// 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 go1.2
package language
import "sort"
var sortStable = sort.Stable

767
vendor/golang.org/x/text/language/index.go generated vendored
View file

@ -1,767 +0,0 @@
// Code generated by running "go generate" in golang.org/x/text. DO NOT EDIT.
package language
// NumCompactTags is the number of common tags. The maximum tag is
// NumCompactTags-1.
const NumCompactTags = 752
var specialTags = []Tag{ // 2 elements
0: {lang: 0xd5, region: 0x6d, script: 0x0, pVariant: 0x5, pExt: 0xe, str: "ca-ES-valencia"},
1: {lang: 0x134, region: 0x134, script: 0x0, pVariant: 0x5, pExt: 0x5, str: "en-US-u-va-posix"},
} // Size: 72 bytes
var coreTags = map[uint32]uint16{
0x0: 0, // und
0x01500000: 3, // af
0x015000d1: 4, // af-NA
0x01500160: 5, // af-ZA
0x01b00000: 6, // agq
0x01b00051: 7, // agq-CM
0x02000000: 8, // ak
0x0200007f: 9, // ak-GH
0x02600000: 10, // am
0x0260006e: 11, // am-ET
0x03900000: 12, // ar
0x03900001: 13, // ar-001
0x03900022: 14, // ar-AE
0x03900038: 15, // ar-BH
0x03900061: 16, // ar-DJ
0x03900066: 17, // ar-DZ
0x0390006a: 18, // ar-EG
0x0390006b: 19, // ar-EH
0x0390006c: 20, // ar-ER
0x03900096: 21, // ar-IL
0x0390009a: 22, // ar-IQ
0x039000a0: 23, // ar-JO
0x039000a7: 24, // ar-KM
0x039000ab: 25, // ar-KW
0x039000af: 26, // ar-LB
0x039000b8: 27, // ar-LY
0x039000b9: 28, // ar-MA
0x039000c8: 29, // ar-MR
0x039000e0: 30, // ar-OM
0x039000ec: 31, // ar-PS
0x039000f2: 32, // ar-QA
0x03900107: 33, // ar-SA
0x0390010a: 34, // ar-SD
0x03900114: 35, // ar-SO
0x03900116: 36, // ar-SS
0x0390011b: 37, // ar-SY
0x0390011f: 38, // ar-TD
0x03900127: 39, // ar-TN
0x0390015d: 40, // ar-YE
0x03f00000: 41, // ars
0x04200000: 42, // as
0x04200098: 43, // as-IN
0x04300000: 44, // asa
0x0430012e: 45, // asa-TZ
0x04700000: 46, // ast
0x0470006d: 47, // ast-ES
0x05700000: 48, // az
0x0571e000: 49, // az-Cyrl
0x0571e031: 50, // az-Cyrl-AZ
0x05752000: 51, // az-Latn
0x05752031: 52, // az-Latn-AZ
0x05d00000: 53, // bas
0x05d00051: 54, // bas-CM
0x07000000: 55, // be
0x07000046: 56, // be-BY
0x07400000: 57, // bem
0x07400161: 58, // bem-ZM
0x07800000: 59, // bez
0x0780012e: 60, // bez-TZ
0x07d00000: 61, // bg
0x07d00037: 62, // bg-BG
0x08100000: 63, // bh
0x09e00000: 64, // bm
0x09e000c2: 65, // bm-ML
0x0a300000: 66, // bn
0x0a300034: 67, // bn-BD
0x0a300098: 68, // bn-IN
0x0a700000: 69, // bo
0x0a700052: 70, // bo-CN
0x0a700098: 71, // bo-IN
0x0b000000: 72, // br
0x0b000077: 73, // br-FR
0x0b300000: 74, // brx
0x0b300098: 75, // brx-IN
0x0b500000: 76, // bs
0x0b51e000: 77, // bs-Cyrl
0x0b51e032: 78, // bs-Cyrl-BA
0x0b552000: 79, // bs-Latn
0x0b552032: 80, // bs-Latn-BA
0x0d500000: 81, // ca
0x0d500021: 82, // ca-AD
0x0d50006d: 83, // ca-ES
0x0d500077: 84, // ca-FR
0x0d50009d: 85, // ca-IT
0x0da00000: 86, // ce
0x0da00105: 87, // ce-RU
0x0dd00000: 88, // cgg
0x0dd00130: 89, // cgg-UG
0x0e300000: 90, // chr
0x0e300134: 91, // chr-US
0x0e700000: 92, // ckb
0x0e70009a: 93, // ckb-IQ
0x0e70009b: 94, // ckb-IR
0x0f600000: 95, // cs
0x0f60005d: 96, // cs-CZ
0x0fa00000: 97, // cu
0x0fa00105: 98, // cu-RU
0x0fc00000: 99, // cy
0x0fc0007a: 100, // cy-GB
0x0fd00000: 101, // da
0x0fd00062: 102, // da-DK
0x0fd00081: 103, // da-GL
0x10400000: 104, // dav
0x104000a3: 105, // dav-KE
0x10900000: 106, // de
0x1090002d: 107, // de-AT
0x10900035: 108, // de-BE
0x1090004d: 109, // de-CH
0x1090005f: 110, // de-DE
0x1090009d: 111, // de-IT
0x109000b1: 112, // de-LI
0x109000b6: 113, // de-LU
0x11300000: 114, // dje
0x113000d3: 115, // dje-NE
0x11b00000: 116, // dsb
0x11b0005f: 117, // dsb-DE
0x12000000: 118, // dua
0x12000051: 119, // dua-CM
0x12400000: 120, // dv
0x12700000: 121, // dyo
0x12700113: 122, // dyo-SN
0x12900000: 123, // dz
0x12900042: 124, // dz-BT
0x12b00000: 125, // ebu
0x12b000a3: 126, // ebu-KE
0x12c00000: 127, // ee
0x12c0007f: 128, // ee-GH
0x12c00121: 129, // ee-TG
0x13100000: 130, // el
0x1310005c: 131, // el-CY
0x13100086: 132, // el-GR
0x13400000: 133, // en
0x13400001: 134, // en-001
0x1340001a: 135, // en-150
0x13400024: 136, // en-AG
0x13400025: 137, // en-AI
0x1340002c: 138, // en-AS
0x1340002d: 139, // en-AT
0x1340002e: 140, // en-AU
0x13400033: 141, // en-BB
0x13400035: 142, // en-BE
0x13400039: 143, // en-BI
0x1340003c: 144, // en-BM
0x13400041: 145, // en-BS
0x13400045: 146, // en-BW
0x13400047: 147, // en-BZ
0x13400048: 148, // en-CA
0x13400049: 149, // en-CC
0x1340004d: 150, // en-CH
0x1340004f: 151, // en-CK
0x13400051: 152, // en-CM
0x1340005b: 153, // en-CX
0x1340005c: 154, // en-CY
0x1340005f: 155, // en-DE
0x13400060: 156, // en-DG
0x13400062: 157, // en-DK
0x13400063: 158, // en-DM
0x1340006c: 159, // en-ER
0x13400071: 160, // en-FI
0x13400072: 161, // en-FJ
0x13400073: 162, // en-FK
0x13400074: 163, // en-FM
0x1340007a: 164, // en-GB
0x1340007b: 165, // en-GD
0x1340007e: 166, // en-GG
0x1340007f: 167, // en-GH
0x13400080: 168, // en-GI
0x13400082: 169, // en-GM
0x13400089: 170, // en-GU
0x1340008b: 171, // en-GY
0x1340008c: 172, // en-HK
0x13400095: 173, // en-IE
0x13400096: 174, // en-IL
0x13400097: 175, // en-IM
0x13400098: 176, // en-IN
0x13400099: 177, // en-IO
0x1340009e: 178, // en-JE
0x1340009f: 179, // en-JM
0x134000a3: 180, // en-KE
0x134000a6: 181, // en-KI
0x134000a8: 182, // en-KN
0x134000ac: 183, // en-KY
0x134000b0: 184, // en-LC
0x134000b3: 185, // en-LR
0x134000b4: 186, // en-LS
0x134000be: 187, // en-MG
0x134000bf: 188, // en-MH
0x134000c5: 189, // en-MO
0x134000c6: 190, // en-MP
0x134000c9: 191, // en-MS
0x134000ca: 192, // en-MT
0x134000cb: 193, // en-MU
0x134000cd: 194, // en-MW
0x134000cf: 195, // en-MY
0x134000d1: 196, // en-NA
0x134000d4: 197, // en-NF
0x134000d5: 198, // en-NG
0x134000d8: 199, // en-NL
0x134000dc: 200, // en-NR
0x134000de: 201, // en-NU
0x134000df: 202, // en-NZ
0x134000e5: 203, // en-PG
0x134000e6: 204, // en-PH
0x134000e7: 205, // en-PK
0x134000ea: 206, // en-PN
0x134000eb: 207, // en-PR
0x134000ef: 208, // en-PW
0x13400106: 209, // en-RW
0x13400108: 210, // en-SB
0x13400109: 211, // en-SC
0x1340010a: 212, // en-SD
0x1340010b: 213, // en-SE
0x1340010c: 214, // en-SG
0x1340010d: 215, // en-SH
0x1340010e: 216, // en-SI
0x13400111: 217, // en-SL
0x13400116: 218, // en-SS
0x1340011a: 219, // en-SX
0x1340011c: 220, // en-SZ
0x1340011e: 221, // en-TC
0x13400124: 222, // en-TK
0x13400128: 223, // en-TO
0x1340012b: 224, // en-TT
0x1340012c: 225, // en-TV
0x1340012e: 226, // en-TZ
0x13400130: 227, // en-UG
0x13400132: 228, // en-UM
0x13400134: 229, // en-US
0x13400138: 230, // en-VC
0x1340013b: 231, // en-VG
0x1340013c: 232, // en-VI
0x1340013e: 233, // en-VU
0x13400141: 234, // en-WS
0x13400160: 235, // en-ZA
0x13400161: 236, // en-ZM
0x13400163: 237, // en-ZW
0x13700000: 238, // eo
0x13700001: 239, // eo-001
0x13900000: 240, // es
0x1390001e: 241, // es-419
0x1390002b: 242, // es-AR
0x1390003e: 243, // es-BO
0x13900040: 244, // es-BR
0x13900050: 245, // es-CL
0x13900053: 246, // es-CO
0x13900055: 247, // es-CR
0x13900058: 248, // es-CU
0x13900064: 249, // es-DO
0x13900067: 250, // es-EA
0x13900068: 251, // es-EC
0x1390006d: 252, // es-ES
0x13900085: 253, // es-GQ
0x13900088: 254, // es-GT
0x1390008e: 255, // es-HN
0x13900093: 256, // es-IC
0x139000ce: 257, // es-MX
0x139000d7: 258, // es-NI
0x139000e1: 259, // es-PA
0x139000e3: 260, // es-PE
0x139000e6: 261, // es-PH
0x139000eb: 262, // es-PR
0x139000f0: 263, // es-PY
0x13900119: 264, // es-SV
0x13900134: 265, // es-US
0x13900135: 266, // es-UY
0x1390013a: 267, // es-VE
0x13b00000: 268, // et
0x13b00069: 269, // et-EE
0x14000000: 270, // eu
0x1400006d: 271, // eu-ES
0x14100000: 272, // ewo
0x14100051: 273, // ewo-CM
0x14300000: 274, // fa
0x14300023: 275, // fa-AF
0x1430009b: 276, // fa-IR
0x14900000: 277, // ff
0x14900051: 278, // ff-CM
0x14900083: 279, // ff-GN
0x149000c8: 280, // ff-MR
0x14900113: 281, // ff-SN
0x14c00000: 282, // fi
0x14c00071: 283, // fi-FI
0x14e00000: 284, // fil
0x14e000e6: 285, // fil-PH
0x15300000: 286, // fo
0x15300062: 287, // fo-DK
0x15300075: 288, // fo-FO
0x15900000: 289, // fr
0x15900035: 290, // fr-BE
0x15900036: 291, // fr-BF
0x15900039: 292, // fr-BI
0x1590003a: 293, // fr-BJ
0x1590003b: 294, // fr-BL
0x15900048: 295, // fr-CA
0x1590004a: 296, // fr-CD
0x1590004b: 297, // fr-CF
0x1590004c: 298, // fr-CG
0x1590004d: 299, // fr-CH
0x1590004e: 300, // fr-CI
0x15900051: 301, // fr-CM
0x15900061: 302, // fr-DJ
0x15900066: 303, // fr-DZ
0x15900077: 304, // fr-FR
0x15900079: 305, // fr-GA
0x1590007d: 306, // fr-GF
0x15900083: 307, // fr-GN
0x15900084: 308, // fr-GP
0x15900085: 309, // fr-GQ
0x15900090: 310, // fr-HT
0x159000a7: 311, // fr-KM
0x159000b6: 312, // fr-LU
0x159000b9: 313, // fr-MA
0x159000ba: 314, // fr-MC
0x159000bd: 315, // fr-MF
0x159000be: 316, // fr-MG
0x159000c2: 317, // fr-ML
0x159000c7: 318, // fr-MQ
0x159000c8: 319, // fr-MR
0x159000cb: 320, // fr-MU
0x159000d2: 321, // fr-NC
0x159000d3: 322, // fr-NE
0x159000e4: 323, // fr-PF
0x159000e9: 324, // fr-PM
0x15900101: 325, // fr-RE
0x15900106: 326, // fr-RW
0x15900109: 327, // fr-SC
0x15900113: 328, // fr-SN
0x1590011b: 329, // fr-SY
0x1590011f: 330, // fr-TD
0x15900121: 331, // fr-TG
0x15900127: 332, // fr-TN
0x1590013e: 333, // fr-VU
0x1590013f: 334, // fr-WF
0x1590015e: 335, // fr-YT
0x16400000: 336, // fur
0x1640009d: 337, // fur-IT
0x16800000: 338, // fy
0x168000d8: 339, // fy-NL
0x16900000: 340, // ga
0x16900095: 341, // ga-IE
0x17800000: 342, // gd
0x1780007a: 343, // gd-GB
0x18a00000: 344, // gl
0x18a0006d: 345, // gl-ES
0x19c00000: 346, // gsw
0x19c0004d: 347, // gsw-CH
0x19c00077: 348, // gsw-FR
0x19c000b1: 349, // gsw-LI
0x19d00000: 350, // gu
0x19d00098: 351, // gu-IN
0x1a200000: 352, // guw
0x1a400000: 353, // guz
0x1a4000a3: 354, // guz-KE
0x1a500000: 355, // gv
0x1a500097: 356, // gv-IM
0x1ad00000: 357, // ha
0x1ad0007f: 358, // ha-GH
0x1ad000d3: 359, // ha-NE
0x1ad000d5: 360, // ha-NG
0x1b100000: 361, // haw
0x1b100134: 362, // haw-US
0x1b500000: 363, // he
0x1b500096: 364, // he-IL
0x1b700000: 365, // hi
0x1b700098: 366, // hi-IN
0x1ca00000: 367, // hr
0x1ca00032: 368, // hr-BA
0x1ca0008f: 369, // hr-HR
0x1cb00000: 370, // hsb
0x1cb0005f: 371, // hsb-DE
0x1ce00000: 372, // hu
0x1ce00091: 373, // hu-HU
0x1d000000: 374, // hy
0x1d000027: 375, // hy-AM
0x1da00000: 376, // id
0x1da00094: 377, // id-ID
0x1df00000: 378, // ig
0x1df000d5: 379, // ig-NG
0x1e200000: 380, // ii
0x1e200052: 381, // ii-CN
0x1f000000: 382, // is
0x1f00009c: 383, // is-IS
0x1f100000: 384, // it
0x1f10004d: 385, // it-CH
0x1f10009d: 386, // it-IT
0x1f100112: 387, // it-SM
0x1f200000: 388, // iu
0x1f800000: 389, // ja
0x1f8000a1: 390, // ja-JP
0x1fb00000: 391, // jbo
0x1ff00000: 392, // jgo
0x1ff00051: 393, // jgo-CM
0x20200000: 394, // jmc
0x2020012e: 395, // jmc-TZ
0x20600000: 396, // jv
0x20800000: 397, // ka
0x2080007c: 398, // ka-GE
0x20a00000: 399, // kab
0x20a00066: 400, // kab-DZ
0x20e00000: 401, // kaj
0x20f00000: 402, // kam
0x20f000a3: 403, // kam-KE
0x21700000: 404, // kcg
0x21b00000: 405, // kde
0x21b0012e: 406, // kde-TZ
0x21f00000: 407, // kea
0x21f00059: 408, // kea-CV
0x22c00000: 409, // khq
0x22c000c2: 410, // khq-ML
0x23100000: 411, // ki
0x231000a3: 412, // ki-KE
0x23a00000: 413, // kk
0x23a000ad: 414, // kk-KZ
0x23c00000: 415, // kkj
0x23c00051: 416, // kkj-CM
0x23d00000: 417, // kl
0x23d00081: 418, // kl-GL
0x23e00000: 419, // kln
0x23e000a3: 420, // kln-KE
0x24200000: 421, // km
0x242000a5: 422, // km-KH
0x24900000: 423, // kn
0x24900098: 424, // kn-IN
0x24b00000: 425, // ko
0x24b000a9: 426, // ko-KP
0x24b000aa: 427, // ko-KR
0x24d00000: 428, // kok
0x24d00098: 429, // kok-IN
0x26100000: 430, // ks
0x26100098: 431, // ks-IN
0x26200000: 432, // ksb
0x2620012e: 433, // ksb-TZ
0x26400000: 434, // ksf
0x26400051: 435, // ksf-CM
0x26500000: 436, // ksh
0x2650005f: 437, // ksh-DE
0x26b00000: 438, // ku
0x27800000: 439, // kw
0x2780007a: 440, // kw-GB
0x28100000: 441, // ky
0x281000a4: 442, // ky-KG
0x28800000: 443, // lag
0x2880012e: 444, // lag-TZ
0x28c00000: 445, // lb
0x28c000b6: 446, // lb-LU
0x29a00000: 447, // lg
0x29a00130: 448, // lg-UG
0x2a600000: 449, // lkt
0x2a600134: 450, // lkt-US
0x2ac00000: 451, // ln
0x2ac00029: 452, // ln-AO
0x2ac0004a: 453, // ln-CD
0x2ac0004b: 454, // ln-CF
0x2ac0004c: 455, // ln-CG
0x2af00000: 456, // lo
0x2af000ae: 457, // lo-LA
0x2b600000: 458, // lrc
0x2b60009a: 459, // lrc-IQ
0x2b60009b: 460, // lrc-IR
0x2b700000: 461, // lt
0x2b7000b5: 462, // lt-LT
0x2b900000: 463, // lu
0x2b90004a: 464, // lu-CD
0x2bb00000: 465, // luo
0x2bb000a3: 466, // luo-KE
0x2bc00000: 467, // luy
0x2bc000a3: 468, // luy-KE
0x2be00000: 469, // lv
0x2be000b7: 470, // lv-LV
0x2c800000: 471, // mas
0x2c8000a3: 472, // mas-KE
0x2c80012e: 473, // mas-TZ
0x2e000000: 474, // mer
0x2e0000a3: 475, // mer-KE
0x2e400000: 476, // mfe
0x2e4000cb: 477, // mfe-MU
0x2e800000: 478, // mg
0x2e8000be: 479, // mg-MG
0x2e900000: 480, // mgh
0x2e9000d0: 481, // mgh-MZ
0x2eb00000: 482, // mgo
0x2eb00051: 483, // mgo-CM
0x2f600000: 484, // mk
0x2f6000c1: 485, // mk-MK
0x2fb00000: 486, // ml
0x2fb00098: 487, // ml-IN
0x30200000: 488, // mn
0x302000c4: 489, // mn-MN
0x31200000: 490, // mr
0x31200098: 491, // mr-IN
0x31600000: 492, // ms
0x3160003d: 493, // ms-BN
0x316000cf: 494, // ms-MY
0x3160010c: 495, // ms-SG
0x31700000: 496, // mt
0x317000ca: 497, // mt-MT
0x31c00000: 498, // mua
0x31c00051: 499, // mua-CM
0x32800000: 500, // my
0x328000c3: 501, // my-MM
0x33100000: 502, // mzn
0x3310009b: 503, // mzn-IR
0x33800000: 504, // nah
0x33c00000: 505, // naq
0x33c000d1: 506, // naq-NA
0x33e00000: 507, // nb
0x33e000d9: 508, // nb-NO
0x33e0010f: 509, // nb-SJ
0x34500000: 510, // nd
0x34500163: 511, // nd-ZW
0x34700000: 512, // nds
0x3470005f: 513, // nds-DE
0x347000d8: 514, // nds-NL
0x34800000: 515, // ne
0x34800098: 516, // ne-IN
0x348000da: 517, // ne-NP
0x35e00000: 518, // nl
0x35e0002f: 519, // nl-AW
0x35e00035: 520, // nl-BE
0x35e0003f: 521, // nl-BQ
0x35e0005a: 522, // nl-CW
0x35e000d8: 523, // nl-NL
0x35e00115: 524, // nl-SR
0x35e0011a: 525, // nl-SX
0x35f00000: 526, // nmg
0x35f00051: 527, // nmg-CM
0x36100000: 528, // nn
0x361000d9: 529, // nn-NO
0x36300000: 530, // nnh
0x36300051: 531, // nnh-CM
0x36600000: 532, // no
0x36c00000: 533, // nqo
0x36d00000: 534, // nr
0x37100000: 535, // nso
0x37700000: 536, // nus
0x37700116: 537, // nus-SS
0x37e00000: 538, // ny
0x38000000: 539, // nyn
0x38000130: 540, // nyn-UG
0x38700000: 541, // om
0x3870006e: 542, // om-ET
0x387000a3: 543, // om-KE
0x38c00000: 544, // or
0x38c00098: 545, // or-IN
0x38f00000: 546, // os
0x38f0007c: 547, // os-GE
0x38f00105: 548, // os-RU
0x39400000: 549, // pa
0x39405000: 550, // pa-Arab
0x394050e7: 551, // pa-Arab-PK
0x3942f000: 552, // pa-Guru
0x3942f098: 553, // pa-Guru-IN
0x39800000: 554, // pap
0x3aa00000: 555, // pl
0x3aa000e8: 556, // pl-PL
0x3b400000: 557, // prg
0x3b400001: 558, // prg-001
0x3b500000: 559, // ps
0x3b500023: 560, // ps-AF
0x3b700000: 561, // pt
0x3b700029: 562, // pt-AO
0x3b700040: 563, // pt-BR
0x3b70004d: 564, // pt-CH
0x3b700059: 565, // pt-CV
0x3b700085: 566, // pt-GQ
0x3b70008a: 567, // pt-GW
0x3b7000b6: 568, // pt-LU
0x3b7000c5: 569, // pt-MO
0x3b7000d0: 570, // pt-MZ
0x3b7000ed: 571, // pt-PT
0x3b700117: 572, // pt-ST
0x3b700125: 573, // pt-TL
0x3bb00000: 574, // qu
0x3bb0003e: 575, // qu-BO
0x3bb00068: 576, // qu-EC
0x3bb000e3: 577, // qu-PE
0x3cb00000: 578, // rm
0x3cb0004d: 579, // rm-CH
0x3d000000: 580, // rn
0x3d000039: 581, // rn-BI
0x3d300000: 582, // ro
0x3d3000bb: 583, // ro-MD
0x3d300103: 584, // ro-RO
0x3d500000: 585, // rof
0x3d50012e: 586, // rof-TZ
0x3d900000: 587, // ru
0x3d900046: 588, // ru-BY
0x3d9000a4: 589, // ru-KG
0x3d9000ad: 590, // ru-KZ
0x3d9000bb: 591, // ru-MD
0x3d900105: 592, // ru-RU
0x3d90012f: 593, // ru-UA
0x3dc00000: 594, // rw
0x3dc00106: 595, // rw-RW
0x3dd00000: 596, // rwk
0x3dd0012e: 597, // rwk-TZ
0x3e200000: 598, // sah
0x3e200105: 599, // sah-RU
0x3e300000: 600, // saq
0x3e3000a3: 601, // saq-KE
0x3e900000: 602, // sbp
0x3e90012e: 603, // sbp-TZ
0x3f200000: 604, // sdh
0x3f300000: 605, // se
0x3f300071: 606, // se-FI
0x3f3000d9: 607, // se-NO
0x3f30010b: 608, // se-SE
0x3f500000: 609, // seh
0x3f5000d0: 610, // seh-MZ
0x3f700000: 611, // ses
0x3f7000c2: 612, // ses-ML
0x3f800000: 613, // sg
0x3f80004b: 614, // sg-CF
0x3fe00000: 615, // shi
0x3fe52000: 616, // shi-Latn
0x3fe520b9: 617, // shi-Latn-MA
0x3fed2000: 618, // shi-Tfng
0x3fed20b9: 619, // shi-Tfng-MA
0x40200000: 620, // si
0x402000b2: 621, // si-LK
0x40800000: 622, // sk
0x40800110: 623, // sk-SK
0x40c00000: 624, // sl
0x40c0010e: 625, // sl-SI
0x41200000: 626, // sma
0x41300000: 627, // smi
0x41400000: 628, // smj
0x41500000: 629, // smn
0x41500071: 630, // smn-FI
0x41800000: 631, // sms
0x41900000: 632, // sn
0x41900163: 633, // sn-ZW
0x41f00000: 634, // so
0x41f00061: 635, // so-DJ
0x41f0006e: 636, // so-ET
0x41f000a3: 637, // so-KE
0x41f00114: 638, // so-SO
0x42700000: 639, // sq
0x42700026: 640, // sq-AL
0x427000c1: 641, // sq-MK
0x4270014c: 642, // sq-XK
0x42800000: 643, // sr
0x4281e000: 644, // sr-Cyrl
0x4281e032: 645, // sr-Cyrl-BA
0x4281e0bc: 646, // sr-Cyrl-ME
0x4281e104: 647, // sr-Cyrl-RS
0x4281e14c: 648, // sr-Cyrl-XK
0x42852000: 649, // sr-Latn
0x42852032: 650, // sr-Latn-BA
0x428520bc: 651, // sr-Latn-ME
0x42852104: 652, // sr-Latn-RS
0x4285214c: 653, // sr-Latn-XK
0x42d00000: 654, // ss
0x43000000: 655, // ssy
0x43100000: 656, // st
0x43a00000: 657, // sv
0x43a00030: 658, // sv-AX
0x43a00071: 659, // sv-FI
0x43a0010b: 660, // sv-SE
0x43b00000: 661, // sw
0x43b0004a: 662, // sw-CD
0x43b000a3: 663, // sw-KE
0x43b0012e: 664, // sw-TZ
0x43b00130: 665, // sw-UG
0x44400000: 666, // syr
0x44600000: 667, // ta
0x44600098: 668, // ta-IN
0x446000b2: 669, // ta-LK
0x446000cf: 670, // ta-MY
0x4460010c: 671, // ta-SG
0x45700000: 672, // te
0x45700098: 673, // te-IN
0x45a00000: 674, // teo
0x45a000a3: 675, // teo-KE
0x45a00130: 676, // teo-UG
0x46100000: 677, // th
0x46100122: 678, // th-TH
0x46500000: 679, // ti
0x4650006c: 680, // ti-ER
0x4650006e: 681, // ti-ET
0x46700000: 682, // tig
0x46c00000: 683, // tk
0x46c00126: 684, // tk-TM
0x47600000: 685, // tn
0x47800000: 686, // to
0x47800128: 687, // to-TO
0x48000000: 688, // tr
0x4800005c: 689, // tr-CY
0x4800012a: 690, // tr-TR
0x48400000: 691, // ts
0x49a00000: 692, // twq
0x49a000d3: 693, // twq-NE
0x49f00000: 694, // tzm
0x49f000b9: 695, // tzm-MA
0x4a200000: 696, // ug
0x4a200052: 697, // ug-CN
0x4a400000: 698, // uk
0x4a40012f: 699, // uk-UA
0x4aa00000: 700, // ur
0x4aa00098: 701, // ur-IN
0x4aa000e7: 702, // ur-PK
0x4b200000: 703, // uz
0x4b205000: 704, // uz-Arab
0x4b205023: 705, // uz-Arab-AF
0x4b21e000: 706, // uz-Cyrl
0x4b21e136: 707, // uz-Cyrl-UZ
0x4b252000: 708, // uz-Latn
0x4b252136: 709, // uz-Latn-UZ
0x4b400000: 710, // vai
0x4b452000: 711, // vai-Latn
0x4b4520b3: 712, // vai-Latn-LR
0x4b4d9000: 713, // vai-Vaii
0x4b4d90b3: 714, // vai-Vaii-LR
0x4b600000: 715, // ve
0x4b900000: 716, // vi
0x4b90013d: 717, // vi-VN
0x4bf00000: 718, // vo
0x4bf00001: 719, // vo-001
0x4c200000: 720, // vun
0x4c20012e: 721, // vun-TZ
0x4c400000: 722, // wa
0x4c500000: 723, // wae
0x4c50004d: 724, // wae-CH
0x4db00000: 725, // wo
0x4e800000: 726, // xh
0x4f100000: 727, // xog
0x4f100130: 728, // xog-UG
0x4ff00000: 729, // yav
0x4ff00051: 730, // yav-CM
0x50800000: 731, // yi
0x50800001: 732, // yi-001
0x50e00000: 733, // yo
0x50e0003a: 734, // yo-BJ
0x50e000d5: 735, // yo-NG
0x51500000: 736, // yue
0x5150008c: 737, // yue-HK
0x51e00000: 738, // zgh
0x51e000b9: 739, // zgh-MA
0x51f00000: 740, // zh
0x51f34000: 741, // zh-Hans
0x51f34052: 742, // zh-Hans-CN
0x51f3408c: 743, // zh-Hans-HK
0x51f340c5: 744, // zh-Hans-MO
0x51f3410c: 745, // zh-Hans-SG
0x51f35000: 746, // zh-Hant
0x51f3508c: 747, // zh-Hant-HK
0x51f350c5: 748, // zh-Hant-MO
0x51f3512d: 749, // zh-Hant-TW
0x52400000: 750, // zu
0x52400160: 751, // zu-ZA
}
// Total table size 4580 bytes (4KiB); checksum: A7F72A2A

975
vendor/golang.org/x/text/language/language.go generated vendored
View file

@ -1,975 +0,0 @@
// 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.
//go:generate go run maketables.go gen_common.go -output tables.go
//go:generate go run gen_index.go
// Package language implements BCP 47 language tags and related functionality.
//
// The Tag type, which is used to represent languages, is agnostic to the
// meaning of its subtags. Tags are not fully canonicalized to preserve
// information that may be valuable in certain contexts. As a consequence, two
// different tags may represent identical languages.
//
// Initializing language- or locale-specific components usually consists of
// two steps. The first step is to select a display language based on the
// preferred languages of the user and the languages supported by an application.
// The second step is to create the language-specific services based on
// this selection. Each is discussed in more details below.
//
// Matching preferred against supported languages
//
// An application may support various languages. This list is typically limited
// by the languages for which there exists translations of the user interface.
// Similarly, a user may provide a list of preferred languages which is limited
// by the languages understood by this user.
// An application should use a Matcher to find the best supported language based
// on the user's preferred list.
// Matchers are aware of the intricacies of equivalence between languages.
// The default Matcher implementation takes into account things such as
// deprecated subtags, legacy tags, and mutual intelligibility between scripts
// and languages.
//
// A Matcher for English, Australian English, Danish, and standard Mandarin can
// be defined as follows:
//
// var matcher = language.NewMatcher([]language.Tag{
// language.English, // The first language is used as fallback.
// language.MustParse("en-AU"),
// language.Danish,
// language.Chinese,
// })
//
// The following code selects the best match for someone speaking Spanish and
// Norwegian:
//
// preferred := []language.Tag{ language.Spanish, language.Norwegian }
// tag, _, _ := matcher.Match(preferred...)
//
// In this case, the best match is Danish, as Danish is sufficiently a match to
// Norwegian to not have to fall back to the default.
// See ParseAcceptLanguage on how to handle the Accept-Language HTTP header.
//
// Selecting language-specific services
//
// One should always use the Tag returned by the Matcher to create an instance
// of any of the language-specific services provided by the text repository.
// This prevents the mixing of languages, such as having a different language for
// messages and display names, as well as improper casing or sorting order for
// the selected language.
// Using the returned Tag also allows user-defined settings, such as collation
// order or numbering system to be transparently passed as options.
//
// If you have language-specific data in your application, however, it will in
// most cases suffice to use the index returned by the matcher to identify
// the user language.
// The following loop provides an alternative in case this is not sufficient:
//
// supported := map[language.Tag]data{
// language.English: enData,
// language.MustParse("en-AU"): enAUData,
// language.Danish: daData,
// language.Chinese: zhData,
// }
// tag, _, _ := matcher.Match(preferred...)
// for ; tag != language.Und; tag = tag.Parent() {
// if v, ok := supported[tag]; ok {
// return v
// }
// }
// return enData // should not reach here
//
// Repeatedly taking the Parent of the tag returned by Match will eventually
// match one of the tags used to initialize the Matcher.
//
// Canonicalization
//
// By default, only legacy and deprecated tags are converted into their
// canonical equivalent. All other information is preserved. This approach makes
// the confidence scores more accurate and allows matchers to distinguish
// between variants that are otherwise lost.
//
// As a consequence, two tags that should be treated as identical according to
// BCP 47 or CLDR, like "en-Latn" and "en", will be represented differently. The
// Matchers will handle such distinctions, though, and are aware of the
// equivalence relations. The CanonType type can be used to alter the
// canonicalization form.
//
// References
//
// BCP 47 - Tags for Identifying Languages
// http://tools.ietf.org/html/bcp47
package language // import "golang.org/x/text/language"
// TODO: Remove above NOTE after:
// - verifying that tables are dropped correctly (most notably matcher tables).
import (
"errors"
"fmt"
"strings"
)
const (
// maxCoreSize is the maximum size of a BCP 47 tag without variants and
// extensions. Equals max lang (3) + script (4) + max reg (3) + 2 dashes.
maxCoreSize = 12
// max99thPercentileSize is a somewhat arbitrary buffer size that presumably
// is large enough to hold at least 99% of the BCP 47 tags.
max99thPercentileSize = 32
// maxSimpleUExtensionSize is the maximum size of a -u extension with one
// key-type pair. Equals len("-u-") + key (2) + dash + max value (8).
maxSimpleUExtensionSize = 14
)
// Tag represents a BCP 47 language tag. It is used to specify an instance of a
// specific language or locale. All language tag values are guaranteed to be
// well-formed.
type Tag struct {
lang langID
region regionID
script scriptID
pVariant byte // offset in str, includes preceding '-'
pExt uint16 // offset of first extension, includes preceding '-'
// str is the string representation of the Tag. It will only be used if the
// tag has variants or extensions.
str string
}
// Make is a convenience wrapper for Parse that omits the error.
// In case of an error, a sensible default is returned.
func Make(s string) Tag {
return Default.Make(s)
}
// Make is a convenience wrapper for c.Parse that omits the error.
// In case of an error, a sensible default is returned.
func (c CanonType) Make(s string) Tag {
t, _ := c.Parse(s)
return t
}
// Raw returns the raw base language, script and region, without making an
// attempt to infer their values.
func (t Tag) Raw() (b Base, s Script, r Region) {
return Base{t.lang}, Script{t.script}, Region{t.region}
}
// equalTags compares language, script and region subtags only.
func (t Tag) equalTags(a Tag) bool {
return t.lang == a.lang && t.script == a.script && t.region == a.region
}
// IsRoot returns true if t is equal to language "und".
func (t Tag) IsRoot() bool {
if int(t.pVariant) < len(t.str) {
return false
}
return t.equalTags(und)
}
// private reports whether the Tag consists solely of a private use tag.
func (t Tag) private() bool {
return t.str != "" && t.pVariant == 0
}
// CanonType can be used to enable or disable various types of canonicalization.
type CanonType int
const (
// Replace deprecated base languages with their preferred replacements.
DeprecatedBase CanonType = 1 << iota
// Replace deprecated scripts with their preferred replacements.
DeprecatedScript
// Replace deprecated regions with their preferred replacements.
DeprecatedRegion
// Remove redundant scripts.
SuppressScript
// Normalize legacy encodings. This includes legacy languages defined in
// CLDR as well as bibliographic codes defined in ISO-639.
Legacy
// Map the dominant language of a macro language group to the macro language
// subtag. For example cmn -> zh.
Macro
// The CLDR flag should be used if full compatibility with CLDR is required.
// There are a few cases where language.Tag may differ from CLDR. To follow all
// of CLDR's suggestions, use All|CLDR.
CLDR
// Raw can be used to Compose or Parse without Canonicalization.
Raw CanonType = 0
// Replace all deprecated tags with their preferred replacements.
Deprecated = DeprecatedBase | DeprecatedScript | DeprecatedRegion
// All canonicalizations recommended by BCP 47.
BCP47 = Deprecated | SuppressScript
// All canonicalizations.
All = BCP47 | Legacy | Macro
// Default is the canonicalization used by Parse, Make and Compose. To
// preserve as much information as possible, canonicalizations that remove
// potentially valuable information are not included. The Matcher is
// designed to recognize similar tags that would be the same if
// they were canonicalized using All.
Default = Deprecated | Legacy
canonLang = DeprecatedBase | Legacy | Macro
// TODO: LikelyScript, LikelyRegion: suppress similar to ICU.
)
// canonicalize returns the canonicalized equivalent of the tag and
// whether there was any change.
func (t Tag) canonicalize(c CanonType) (Tag, bool) {
if c == Raw {
return t, false
}
changed := false
if c&SuppressScript != 0 {
if t.lang < langNoIndexOffset && uint8(t.script) == suppressScript[t.lang] {
t.script = 0
changed = true
}
}
if c&canonLang != 0 {
for {
if l, aliasType := normLang(t.lang); l != t.lang {
switch aliasType {
case langLegacy:
if c&Legacy != 0 {
if t.lang == _sh && t.script == 0 {
t.script = _Latn
}
t.lang = l
changed = true
}
case langMacro:
if c&Macro != 0 {
// We deviate here from CLDR. The mapping "nb" -> "no"
// qualifies as a typical Macro language mapping. However,
// for legacy reasons, CLDR maps "no", the macro language
// code for Norwegian, to the dominant variant "nb". This
// change is currently under consideration for CLDR as well.
// See http://unicode.org/cldr/trac/ticket/2698 and also
// http://unicode.org/cldr/trac/ticket/1790 for some of the
// practical implications. TODO: this check could be removed
// if CLDR adopts this change.
if c&CLDR == 0 || t.lang != _nb {
changed = true
t.lang = l
}
}
case langDeprecated:
if c&DeprecatedBase != 0 {
if t.lang == _mo && t.region == 0 {
t.region = _MD
}
t.lang = l
changed = true
// Other canonicalization types may still apply.
continue
}
}
} else if c&Legacy != 0 && t.lang == _no && c&CLDR != 0 {
t.lang = _nb
changed = true
}
break
}
}
if c&DeprecatedScript != 0 {
if t.script == _Qaai {
changed = true
t.script = _Zinh
}
}
if c&DeprecatedRegion != 0 {
if r := normRegion(t.region); r != 0 {
changed = true
t.region = r
}
}
return t, changed
}
// Canonicalize returns the canonicalized equivalent of the tag.
func (c CanonType) Canonicalize(t Tag) (Tag, error) {
t, changed := t.canonicalize(c)
if changed {
t.remakeString()
}
return t, nil
}
// Confidence indicates the level of certainty for a given return value.
// For example, Serbian may be written in Cyrillic or Latin script.
// The confidence level indicates whether a value was explicitly specified,
// whether it is typically the only possible value, or whether there is
// an ambiguity.
type Confidence int
const (
No Confidence = iota // full confidence that there was no match
Low // most likely value picked out of a set of alternatives
High // value is generally assumed to be the correct match
Exact // exact match or explicitly specified value
)
var confName = []string{"No", "Low", "High", "Exact"}
func (c Confidence) String() string {
return confName[c]
}
// remakeString is used to update t.str in case lang, script or region changed.
// It is assumed that pExt and pVariant still point to the start of the
// respective parts.
func (t *Tag) remakeString() {
if t.str == "" {
return
}
extra := t.str[t.pVariant:]
if t.pVariant > 0 {
extra = extra[1:]
}
if t.equalTags(und) && strings.HasPrefix(extra, "x-") {
t.str = extra
t.pVariant = 0
t.pExt = 0
return
}
var buf [max99thPercentileSize]byte // avoid extra memory allocation in most cases.
b := buf[:t.genCoreBytes(buf[:])]
if extra != "" {
diff := len(b) - int(t.pVariant)
b = append(b, '-')
b = append(b, extra...)
t.pVariant = uint8(int(t.pVariant) + diff)
t.pExt = uint16(int(t.pExt) + diff)
} else {
t.pVariant = uint8(len(b))
t.pExt = uint16(len(b))
}
t.str = string(b)
}
// genCoreBytes writes a string for the base languages, script and region tags
// to the given buffer and returns the number of bytes written. It will never
// write more than maxCoreSize bytes.
func (t *Tag) genCoreBytes(buf []byte) int {
n := t.lang.stringToBuf(buf[:])
if t.script != 0 {
n += copy(buf[n:], "-")
n += copy(buf[n:], t.script.String())
}
if t.region != 0 {
n += copy(buf[n:], "-")
n += copy(buf[n:], t.region.String())
}
return n
}
// String returns the canonical string representation of the language tag.
func (t Tag) String() string {
if t.str != "" {
return t.str
}
if t.script == 0 && t.region == 0 {
return t.lang.String()
}
buf := [maxCoreSize]byte{}
return string(buf[:t.genCoreBytes(buf[:])])
}
// Base returns the base language of the language tag. If the base language is
// unspecified, an attempt will be made to infer it from the context.
// It uses a variant of CLDR's Add Likely Subtags algorithm. This is subject to change.
func (t Tag) Base() (Base, Confidence) {
if t.lang != 0 {
return Base{t.lang}, Exact
}
c := High
if t.script == 0 && !(Region{t.region}).IsCountry() {
c = Low
}
if tag, err := addTags(t); err == nil && tag.lang != 0 {
return Base{tag.lang}, c
}
return Base{0}, No
}
// Script infers the script for the language tag. If it was not explicitly given, it will infer
// a most likely candidate.
// If more than one script is commonly used for a language, the most likely one
// is returned with a low confidence indication. For example, it returns (Cyrl, Low)
// for Serbian.
// If a script cannot be inferred (Zzzz, No) is returned. We do not use Zyyy (undetermined)
// as one would suspect from the IANA registry for BCP 47. In a Unicode context Zyyy marks
// common characters (like 1, 2, 3, '.', etc.) and is therefore more like multiple scripts.
// See http://www.unicode.org/reports/tr24/#Values for more details. Zzzz is also used for
// unknown value in CLDR. (Zzzz, Exact) is returned if Zzzz was explicitly specified.
// Note that an inferred script is never guaranteed to be the correct one. Latin is
// almost exclusively used for Afrikaans, but Arabic has been used for some texts
// in the past. Also, the script that is commonly used may change over time.
// It uses a variant of CLDR's Add Likely Subtags algorithm. This is subject to change.
func (t Tag) Script() (Script, Confidence) {
if t.script != 0 {
return Script{t.script}, Exact
}
sc, c := scriptID(_Zzzz), No
if t.lang < langNoIndexOffset {
if scr := scriptID(suppressScript[t.lang]); scr != 0 {
// Note: it is not always the case that a language with a suppress
// script value is only written in one script (e.g. kk, ms, pa).
if t.region == 0 {
return Script{scriptID(scr)}, High
}
sc, c = scr, High
}
}
if tag, err := addTags(t); err == nil {
if tag.script != sc {
sc, c = tag.script, Low
}
} else {
t, _ = (Deprecated | Macro).Canonicalize(t)
if tag, err := addTags(t); err == nil && tag.script != sc {
sc, c = tag.script, Low
}
}
return Script{sc}, c
}
// Region returns the region for the language tag. If it was not explicitly given, it will
// infer a most likely candidate from the context.
// It uses a variant of CLDR's Add Likely Subtags algorithm. This is subject to change.
func (t Tag) Region() (Region, Confidence) {
if t.region != 0 {
return Region{t.region}, Exact
}
if t, err := addTags(t); err == nil {
return Region{t.region}, Low // TODO: differentiate between high and low.
}
t, _ = (Deprecated | Macro).Canonicalize(t)
if tag, err := addTags(t); err == nil {
return Region{tag.region}, Low
}
return Region{_ZZ}, No // TODO: return world instead of undetermined?
}
// Variant returns the variants specified explicitly for this language tag.
// or nil if no variant was specified.
func (t Tag) Variants() []Variant {
v := []Variant{}
if int(t.pVariant) < int(t.pExt) {
for x, str := "", t.str[t.pVariant:t.pExt]; str != ""; {
x, str = nextToken(str)
v = append(v, Variant{x})
}
}
return v
}
// Parent returns the CLDR parent of t. In CLDR, missing fields in data for a
// specific language are substituted with fields from the parent language.
// The parent for a language may change for newer versions of CLDR.
func (t Tag) Parent() Tag {
if t.str != "" {
// Strip the variants and extensions.
t, _ = Raw.Compose(t.Raw())
if t.region == 0 && t.script != 0 && t.lang != 0 {
base, _ := addTags(Tag{lang: t.lang})
if base.script == t.script {
return Tag{lang: t.lang}
}
}
return t
}
if t.lang != 0 {
if t.region != 0 {
maxScript := t.script
if maxScript == 0 {
max, _ := addTags(t)
maxScript = max.script
}
for i := range parents {
if langID(parents[i].lang) == t.lang && scriptID(parents[i].maxScript) == maxScript {
for _, r := range parents[i].fromRegion {
if regionID(r) == t.region {
return Tag{
lang: t.lang,
script: scriptID(parents[i].script),
region: regionID(parents[i].toRegion),
}
}
}
}
}
// Strip the script if it is the default one.
base, _ := addTags(Tag{lang: t.lang})
if base.script != maxScript {
return Tag{lang: t.lang, script: maxScript}
}
return Tag{lang: t.lang}
} else if t.script != 0 {
// The parent for an base-script pair with a non-default script is
// "und" instead of the base language.
base, _ := addTags(Tag{lang: t.lang})
if base.script != t.script {
return und
}
return Tag{lang: t.lang}
}
}
return und
}
// returns token t and the rest of the string.
func nextToken(s string) (t, tail string) {
p := strings.Index(s[1:], "-")
if p == -1 {
return s[1:], ""
}
p++
return s[1:p], s[p:]
}
// Extension is a single BCP 47 extension.
type Extension struct {
s string
}
// String returns the string representation of the extension, including the
// type tag.
func (e Extension) String() string {
return e.s
}
// ParseExtension parses s as an extension and returns it on success.
func ParseExtension(s string) (e Extension, err error) {
scan := makeScannerString(s)
var end int
if n := len(scan.token); n != 1 {
return Extension{}, errSyntax
}
scan.toLower(0, len(scan.b))
end = parseExtension(&scan)
if end != len(s) {
return Extension{}, errSyntax
}
return Extension{string(scan.b)}, nil
}
// Type returns the one-byte extension type of e. It returns 0 for the zero
// exception.
func (e Extension) Type() byte {
if e.s == "" {
return 0
}
return e.s[0]
}
// Tokens returns the list of tokens of e.
func (e Extension) Tokens() []string {
return strings.Split(e.s, "-")
}
// Extension returns the extension of type x for tag t. It will return
// false for ok if t does not have the requested extension. The returned
// extension will be invalid in this case.
func (t Tag) Extension(x byte) (ext Extension, ok bool) {
for i := int(t.pExt); i < len(t.str)-1; {
var ext string
i, ext = getExtension(t.str, i)
if ext[0] == x {
return Extension{ext}, true
}
}
return Extension{}, false
}
// Extensions returns all extensions of t.
func (t Tag) Extensions() []Extension {
e := []Extension{}
for i := int(t.pExt); i < len(t.str)-1; {
var ext string
i, ext = getExtension(t.str, i)
e = append(e, Extension{ext})
}
return e
}
// TypeForKey returns the type associated with the given key, where key and type
// are of the allowed values defined for the Unicode locale extension ('u') in
// http://www.unicode.org/reports/tr35/#Unicode_Language_and_Locale_Identifiers.
// TypeForKey will traverse the inheritance chain to get the correct value.
func (t Tag) TypeForKey(key string) string {
if start, end, _ := t.findTypeForKey(key); end != start {
return t.str[start:end]
}
return ""
}
var (
errPrivateUse = errors.New("cannot set a key on a private use tag")
errInvalidArguments = errors.New("invalid key or type")
)
// SetTypeForKey returns a new Tag with the key set to type, where key and type
// are of the allowed values defined for the Unicode locale extension ('u') in
// http://www.unicode.org/reports/tr35/#Unicode_Language_and_Locale_Identifiers.
// An empty value removes an existing pair with the same key.
func (t Tag) SetTypeForKey(key, value string) (Tag, error) {
if t.private() {
return t, errPrivateUse
}
if len(key) != 2 {
return t, errInvalidArguments
}
// Remove the setting if value is "".
if value == "" {
start, end, _ := t.findTypeForKey(key)
if start != end {
// Remove key tag and leading '-'.
start -= 4
// Remove a possible empty extension.
if (end == len(t.str) || t.str[end+2] == '-') && t.str[start-2] == '-' {
start -= 2
}
if start == int(t.pVariant) && end == len(t.str) {
t.str = ""
t.pVariant, t.pExt = 0, 0
} else {
t.str = fmt.Sprintf("%s%s", t.str[:start], t.str[end:])
}
}
return t, nil
}
if len(value) < 3 || len(value) > 8 {
return t, errInvalidArguments
}
var (
buf [maxCoreSize + maxSimpleUExtensionSize]byte
uStart int // start of the -u extension.
)
// Generate the tag string if needed.
if t.str == "" {
uStart = t.genCoreBytes(buf[:])
buf[uStart] = '-'
uStart++
}
// Create new key-type pair and parse it to verify.
b := buf[uStart:]
copy(b, "u-")
copy(b[2:], key)
b[4] = '-'
b = b[:5+copy(b[5:], value)]
scan := makeScanner(b)
if parseExtensions(&scan); scan.err != nil {
return t, scan.err
}
// Assemble the replacement string.
if t.str == "" {
t.pVariant, t.pExt = byte(uStart-1), uint16(uStart-1)
t.str = string(buf[:uStart+len(b)])
} else {
s := t.str
start, end, hasExt := t.findTypeForKey(key)
if start == end {
if hasExt {
b = b[2:]
}
t.str = fmt.Sprintf("%s-%s%s", s[:start], b, s[end:])
} else {
t.str = fmt.Sprintf("%s%s%s", s[:start], value, s[end:])
}
}
return t, nil
}
// findKeyAndType returns the start and end position for the type corresponding
// to key or the point at which to insert the key-value pair if the type
// wasn't found. The hasExt return value reports whether an -u extension was present.
// Note: the extensions are typically very small and are likely to contain
// only one key-type pair.
func (t Tag) findTypeForKey(key string) (start, end int, hasExt bool) {
p := int(t.pExt)
if len(key) != 2 || p == len(t.str) || p == 0 {
return p, p, false
}
s := t.str
// Find the correct extension.
for p++; s[p] != 'u'; p++ {
if s[p] > 'u' {
p--
return p, p, false
}
if p = nextExtension(s, p); p == len(s) {
return len(s), len(s), false
}
}
// Proceed to the hyphen following the extension name.
p++
// curKey is the key currently being processed.
curKey := ""
// Iterate over keys until we get the end of a section.
for {
// p points to the hyphen preceding the current token.
if p3 := p + 3; s[p3] == '-' {
// Found a key.
// Check whether we just processed the key that was requested.
if curKey == key {
return start, p, true
}
// Set to the next key and continue scanning type tokens.
curKey = s[p+1 : p3]
if curKey > key {
return p, p, true
}
// Start of the type token sequence.
start = p + 4
// A type is at least 3 characters long.
p += 7 // 4 + 3
} else {
// Attribute or type, which is at least 3 characters long.
p += 4
}
// p points past the third character of a type or attribute.
max := p + 5 // maximum length of token plus hyphen.
if len(s) < max {
max = len(s)
}
for ; p < max && s[p] != '-'; p++ {
}
// Bail if we have exhausted all tokens or if the next token starts
// a new extension.
if p == len(s) || s[p+2] == '-' {
if curKey == key {
return start, p, true
}
return p, p, true
}
}
}
// CompactIndex returns an index, where 0 <= index < NumCompactTags, for tags
// for which data exists in the text repository. The index will change over time
// and should not be stored in persistent storage. Extensions, except for the
// 'va' type of the 'u' extension, are ignored. It will return 0, false if no
// compact tag exists, where 0 is the index for the root language (Und).
func CompactIndex(t Tag) (index int, ok bool) {
// TODO: perhaps give more frequent tags a lower index.
// TODO: we could make the indexes stable. This will excluded some
// possibilities for optimization, so don't do this quite yet.
b, s, r := t.Raw()
if len(t.str) > 0 {
if strings.HasPrefix(t.str, "x-") {
// We have no entries for user-defined tags.
return 0, false
}
if uint16(t.pVariant) != t.pExt {
// There are no tags with variants and an u-va type.
if t.TypeForKey("va") != "" {
return 0, false
}
t, _ = Raw.Compose(b, s, r, t.Variants())
} else if _, ok := t.Extension('u'); ok {
// Strip all but the 'va' entry.
variant := t.TypeForKey("va")
t, _ = Raw.Compose(b, s, r)
t, _ = t.SetTypeForKey("va", variant)
}
if len(t.str) > 0 {
// We have some variants.
for i, s := range specialTags {
if s == t {
return i + 1, true
}
}
return 0, false
}
}
// No variants specified: just compare core components.
// The key has the form lllssrrr, where l, s, and r are nibbles for
// respectively the langID, scriptID, and regionID.
key := uint32(b.langID) << (8 + 12)
key |= uint32(s.scriptID) << 12
key |= uint32(r.regionID)
x, ok := coreTags[key]
return int(x), ok
}
// Base is an ISO 639 language code, used for encoding the base language
// of a language tag.
type Base struct {
langID
}
// ParseBase parses a 2- or 3-letter ISO 639 code.
// It returns a ValueError if s is a well-formed but unknown language identifier
// or another error if another error occurred.
func ParseBase(s string) (Base, error) {
if n := len(s); n < 2 || 3 < n {
return Base{}, errSyntax
}
var buf [3]byte
l, err := getLangID(buf[:copy(buf[:], s)])
return Base{l}, err
}
// Script is a 4-letter ISO 15924 code for representing scripts.
// It is idiomatically represented in title case.
type Script struct {
scriptID
}
// ParseScript parses a 4-letter ISO 15924 code.
// It returns a ValueError if s is a well-formed but unknown script identifier
// or another error if another error occurred.
func ParseScript(s string) (Script, error) {
if len(s) != 4 {
return Script{}, errSyntax
}
var buf [4]byte
sc, err := getScriptID(script, buf[:copy(buf[:], s)])
return Script{sc}, err
}
// Region is an ISO 3166-1 or UN M.49 code for representing countries and regions.
type Region struct {
regionID
}
// EncodeM49 returns the Region for the given UN M.49 code.
// It returns an error if r is not a valid code.
func EncodeM49(r int) (Region, error) {
rid, err := getRegionM49(r)
return Region{rid}, err
}
// ParseRegion parses a 2- or 3-letter ISO 3166-1 or a UN M.49 code.
// It returns a ValueError if s is a well-formed but unknown region identifier
// or another error if another error occurred.
func ParseRegion(s string) (Region, error) {
if n := len(s); n < 2 || 3 < n {
return Region{}, errSyntax
}
var buf [3]byte
r, err := getRegionID(buf[:copy(buf[:], s)])
return Region{r}, err
}
// IsCountry returns whether this region is a country or autonomous area. This
// includes non-standard definitions from CLDR.
func (r Region) IsCountry() bool {
if r.regionID == 0 || r.IsGroup() || r.IsPrivateUse() && r.regionID != _XK {
return false
}
return true
}
// IsGroup returns whether this region defines a collection of regions. This
// includes non-standard definitions from CLDR.
func (r Region) IsGroup() bool {
if r.regionID == 0 {
return false
}
return int(regionInclusion[r.regionID]) < len(regionContainment)
}
// Contains returns whether Region c is contained by Region r. It returns true
// if c == r.
func (r Region) Contains(c Region) bool {
return r.regionID.contains(c.regionID)
}
func (r regionID) contains(c regionID) bool {
if r == c {
return true
}
g := regionInclusion[r]
if g >= nRegionGroups {
return false
}
m := regionContainment[g]
d := regionInclusion[c]
b := regionInclusionBits[d]
// A contained country may belong to multiple disjoint groups. Matching any
// of these indicates containment. If the contained region is a group, it
// must strictly be a subset.
if d >= nRegionGroups {
return b&m != 0
}
return b&^m == 0
}
var errNoTLD = errors.New("language: region is not a valid ccTLD")
// TLD returns the country code top-level domain (ccTLD). UK is returned for GB.
// In all other cases it returns either the region itself or an error.
//
// This method may return an error for a region for which there exists a
// canonical form with a ccTLD. To get that ccTLD canonicalize r first. The
// region will already be canonicalized it was obtained from a Tag that was
// obtained using any of the default methods.
func (r Region) TLD() (Region, error) {
// See http://en.wikipedia.org/wiki/Country_code_top-level_domain for the
// difference between ISO 3166-1 and IANA ccTLD.
if r.regionID == _GB {
r = Region{_UK}
}
if (r.typ() & ccTLD) == 0 {
return Region{}, errNoTLD
}
return r, nil
}
// Canonicalize returns the region or a possible replacement if the region is
// deprecated. It will not return a replacement for deprecated regions that
// are split into multiple regions.
func (r Region) Canonicalize() Region {
if cr := normRegion(r.regionID); cr != 0 {
return Region{cr}
}
return r
}
// Variant represents a registered variant of a language as defined by BCP 47.
type Variant struct {
variant string
}
// ParseVariant parses and returns a Variant. An error is returned if s is not
// a valid variant.
func ParseVariant(s string) (Variant, error) {
s = strings.ToLower(s)
if _, ok := variantIndex[s]; ok {
return Variant{s}, nil
}
return Variant{}, mkErrInvalid([]byte(s))
}
// String returns the string representation of the variant.
func (v Variant) String() string {
return v.variant
}

396
vendor/golang.org/x/text/language/lookup.go generated vendored
View file

@ -1,396 +0,0 @@
// 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.
package language
import (
"bytes"
"fmt"
"sort"
"strconv"
"golang.org/x/text/internal/tag"
)
// findIndex tries to find the given tag in idx and returns a standardized error
// if it could not be found.
func findIndex(idx tag.Index, key []byte, form string) (index int, err error) {
if !tag.FixCase(form, key) {
return 0, errSyntax
}
i := idx.Index(key)
if i == -1 {
return 0, mkErrInvalid(key)
}
return i, nil
}
func searchUint(imap []uint16, key uint16) int {
return sort.Search(len(imap), func(i int) bool {
return imap[i] >= key
})
}
type langID uint16
// getLangID returns the langID of s if s is a canonical subtag
// or langUnknown if s is not a canonical subtag.
func getLangID(s []byte) (langID, error) {
if len(s) == 2 {
return getLangISO2(s)
}
return getLangISO3(s)
}
// mapLang returns the mapped langID of id according to mapping m.
func normLang(id langID) (langID, langAliasType) {
k := sort.Search(len(langAliasMap), func(i int) bool {
return langAliasMap[i].from >= uint16(id)
})
if k < len(langAliasMap) && langAliasMap[k].from == uint16(id) {
return langID(langAliasMap[k].to), langAliasTypes[k]
}
return id, langAliasTypeUnknown
}
// getLangISO2 returns the langID for the given 2-letter ISO language code
// or unknownLang if this does not exist.
func getLangISO2(s []byte) (langID, error) {
if !tag.FixCase("zz", s) {
return 0, errSyntax
}
if i := lang.Index(s); i != -1 && lang.Elem(i)[3] != 0 {
return langID(i), nil
}
return 0, mkErrInvalid(s)
}
const base = 'z' - 'a' + 1
func strToInt(s []byte) 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
}
}
// getLangISO3 returns the langID for the given 3-letter ISO language code
// or unknownLang if this does not exist.
func getLangISO3(s []byte) (langID, error) {
if tag.FixCase("und", s) {
// first try to match canonical 3-letter entries
for i := lang.Index(s[:2]); i != -1; i = lang.Next(s[:2], i) {
if e := lang.Elem(i); e[3] == 0 && e[2] == s[2] {
// We treat "und" as special and always translate it to "unspecified".
// Note that ZZ and Zzzz are private use and are not treated as
// unspecified by default.
id := langID(i)
if id == nonCanonicalUnd {
return 0, nil
}
return id, nil
}
}
if i := altLangISO3.Index(s); i != -1 {
return langID(altLangIndex[altLangISO3.Elem(i)[3]]), nil
}
n := strToInt(s)
if langNoIndex[n/8]&(1<<(n%8)) != 0 {
return langID(n) + langNoIndexOffset, nil
}
// Check for non-canonical uses of ISO3.
for i := lang.Index(s[:1]); i != -1; i = lang.Next(s[:1], i) {
if e := lang.Elem(i); e[2] == s[1] && e[3] == s[2] {
return langID(i), nil
}
}
return 0, mkErrInvalid(s)
}
return 0, errSyntax
}
// stringToBuf writes the string to b and returns the number of bytes
// written. cap(b) must be >= 3.
func (id langID) stringToBuf(b []byte) int {
if id >= langNoIndexOffset {
intToStr(uint(id)-langNoIndexOffset, b[:3])
return 3
} else if id == 0 {
return copy(b, "und")
}
l := lang[id<<2:]
if l[3] == 0 {
return copy(b, l[:3])
}
return copy(b, l[:2])
}
// String returns the BCP 47 representation of the langID.
// Use b as variable name, instead of id, to ensure the variable
// used is consistent with that of Base in which this type is embedded.
func (b langID) String() string {
if b == 0 {
return "und"
} else if b >= langNoIndexOffset {
b -= langNoIndexOffset
buf := [3]byte{}
intToStr(uint(b), buf[:])
return string(buf[:])
}
l := lang.Elem(int(b))
if l[3] == 0 {
return l[:3]
}
return l[:2]
}
// ISO3 returns the ISO 639-3 language code.
func (b langID) ISO3() string {
if b == 0 || b >= langNoIndexOffset {
return b.String()
}
l := lang.Elem(int(b))
if l[3] == 0 {
return l[:3]
} else if l[2] == 0 {
return altLangISO3.Elem(int(l[3]))[:3]
}
// This allocation will only happen for 3-letter ISO codes
// that are non-canonical BCP 47 language identifiers.
return l[0:1] + l[2:4]
}
// IsPrivateUse reports whether this language code is reserved for private use.
func (b langID) IsPrivateUse() bool {
return langPrivateStart <= b && b <= langPrivateEnd
}
type regionID uint16
// getRegionID returns the region id for s if s is a valid 2-letter region code
// or unknownRegion.
func getRegionID(s []byte) (regionID, error) {
if len(s) == 3 {
if isAlpha(s[0]) {
return getRegionISO3(s)
}
if i, err := strconv.ParseUint(string(s), 10, 10); err == nil {
return getRegionM49(int(i))
}
}
return getRegionISO2(s)
}
// getRegionISO2 returns the regionID for the given 2-letter ISO country code
// or unknownRegion if this does not exist.
func getRegionISO2(s []byte) (regionID, error) {
i, err := findIndex(regionISO, s, "ZZ")
if err != nil {
return 0, err
}
return regionID(i) + isoRegionOffset, nil
}
// getRegionISO3 returns the regionID for the given 3-letter ISO country code
// or unknownRegion if this does not exist.
func getRegionISO3(s []byte) (regionID, error) {
if tag.FixCase("ZZZ", s) {
for i := regionISO.Index(s[:1]); i != -1; i = regionISO.Next(s[:1], i) {
if e := regionISO.Elem(i); e[2] == s[1] && e[3] == s[2] {
return regionID(i) + isoRegionOffset, nil
}
}
for i := 0; i < len(altRegionISO3); i += 3 {
if tag.Compare(altRegionISO3[i:i+3], s) == 0 {
return regionID(altRegionIDs[i/3]), nil
}
}
return 0, mkErrInvalid(s)
}
return 0, errSyntax
}
func getRegionM49(n int) (regionID, error) {
if 0 < n && n <= 999 {
const (
searchBits = 7
regionBits = 9
regionMask = 1<<regionBits - 1
)
idx := n >> searchBits
buf := fromM49[m49Index[idx]:m49Index[idx+1]]
val := uint16(n) << regionBits // we rely on bits shifting out
i := sort.Search(len(buf), func(i int) bool {
return buf[i] >= val
})
if r := fromM49[int(m49Index[idx])+i]; r&^regionMask == val {
return regionID(r & regionMask), nil
}
}
var e ValueError
fmt.Fprint(bytes.NewBuffer([]byte(e.v[:])), n)
return 0, e
}
// normRegion returns a region if r is deprecated or 0 otherwise.
// TODO: consider supporting BYS (-> BLR), CSK (-> 200 or CZ), PHI (-> PHL) and AFI (-> DJ).
// TODO: consider mapping split up regions to new most populous one (like CLDR).
func normRegion(r regionID) regionID {
m := regionOldMap
k := sort.Search(len(m), func(i int) bool {
return m[i].from >= uint16(r)
})
if k < len(m) && m[k].from == uint16(r) {
return regionID(m[k].to)
}
return 0
}
const (
iso3166UserAssigned = 1 << iota
ccTLD
bcp47Region
)
func (r regionID) typ() byte {
return regionTypes[r]
}
// String returns the BCP 47 representation for the region.
// It returns "ZZ" for an unspecified region.
func (r regionID) String() string {
if r < isoRegionOffset {
if r == 0 {
return "ZZ"
}
return fmt.Sprintf("%03d", r.M49())
}
r -= isoRegionOffset
return regionISO.Elem(int(r))[:2]
}
// ISO3 returns the 3-letter ISO code of r.
// Note that not all regions have a 3-letter ISO code.
// In such cases this method returns "ZZZ".
func (r regionID) ISO3() string {
if r < isoRegionOffset {
return "ZZZ"
}
r -= isoRegionOffset
reg := regionISO.Elem(int(r))
switch reg[2] {
case 0:
return altRegionISO3[reg[3]:][:3]
case ' ':
return "ZZZ"
}
return reg[0:1] + reg[2:4]
}
// M49 returns the UN M.49 encoding of r, or 0 if this encoding
// is not defined for r.
func (r regionID) M49() int {
return int(m49[r])
}
// IsPrivateUse reports whether r has the ISO 3166 User-assigned status. This
// may include private-use tags that are assigned by CLDR and used in this
// implementation. So IsPrivateUse and IsCountry can be simultaneously true.
func (r regionID) IsPrivateUse() bool {
return r.typ()&iso3166UserAssigned != 0
}
type scriptID uint8
// getScriptID returns the script id for string s. It assumes that s
// is of the format [A-Z][a-z]{3}.
func getScriptID(idx tag.Index, s []byte) (scriptID, error) {
i, err := findIndex(idx, s, "Zzzz")
return scriptID(i), err
}
// String returns the script code in title case.
// It returns "Zzzz" for an unspecified script.
func (s scriptID) String() string {
if s == 0 {
return "Zzzz"
}
return script.Elem(int(s))
}
// IsPrivateUse reports whether this script code is reserved for private use.
func (s scriptID) IsPrivateUse() bool {
return _Qaaa <= s && s <= _Qabx
}
const (
maxAltTaglen = len("en-US-POSIX")
maxLen = maxAltTaglen
)
var (
// grandfatheredMap holds a mapping from legacy and grandfathered tags to
// their base language or index to more elaborate tag.
grandfatheredMap = map[[maxLen]byte]int16{
[maxLen]byte{'a', 'r', 't', '-', 'l', 'o', 'j', 'b', 'a', 'n'}: _jbo, // art-lojban
[maxLen]byte{'i', '-', 'a', 'm', 'i'}: _ami, // i-ami
[maxLen]byte{'i', '-', 'b', 'n', 'n'}: _bnn, // i-bnn
[maxLen]byte{'i', '-', 'h', 'a', 'k'}: _hak, // i-hak
[maxLen]byte{'i', '-', 'k', 'l', 'i', 'n', 'g', 'o', 'n'}: _tlh, // i-klingon
[maxLen]byte{'i', '-', 'l', 'u', 'x'}: _lb, // i-lux
[maxLen]byte{'i', '-', 'n', 'a', 'v', 'a', 'j', 'o'}: _nv, // i-navajo
[maxLen]byte{'i', '-', 'p', 'w', 'n'}: _pwn, // i-pwn
[maxLen]byte{'i', '-', 't', 'a', 'o'}: _tao, // i-tao
[maxLen]byte{'i', '-', 't', 'a', 'y'}: _tay, // i-tay
[maxLen]byte{'i', '-', 't', 's', 'u'}: _tsu, // i-tsu
[maxLen]byte{'n', 'o', '-', 'b', 'o', 'k'}: _nb, // no-bok
[maxLen]byte{'n', 'o', '-', 'n', 'y', 'n'}: _nn, // no-nyn
[maxLen]byte{'s', 'g', 'n', '-', 'b', 'e', '-', 'f', 'r'}: _sfb, // sgn-BE-FR
[maxLen]byte{'s', 'g', 'n', '-', 'b', 'e', '-', 'n', 'l'}: _vgt, // sgn-BE-NL
[maxLen]byte{'s', 'g', 'n', '-', 'c', 'h', '-', 'd', 'e'}: _sgg, // sgn-CH-DE
[maxLen]byte{'z', 'h', '-', 'g', 'u', 'o', 'y', 'u'}: _cmn, // zh-guoyu
[maxLen]byte{'z', 'h', '-', 'h', 'a', 'k', 'k', 'a'}: _hak, // zh-hakka
[maxLen]byte{'z', 'h', '-', 'm', 'i', 'n', '-', 'n', 'a', 'n'}: _nan, // zh-min-nan
[maxLen]byte{'z', 'h', '-', 'x', 'i', 'a', 'n', 'g'}: _hsn, // zh-xiang
// Grandfathered tags with no modern replacement will be converted as
// follows:
[maxLen]byte{'c', 'e', 'l', '-', 'g', 'a', 'u', 'l', 'i', 's', 'h'}: -1, // cel-gaulish
[maxLen]byte{'e', 'n', '-', 'g', 'b', '-', 'o', 'e', 'd'}: -2, // en-GB-oed
[maxLen]byte{'i', '-', 'd', 'e', 'f', 'a', 'u', 'l', 't'}: -3, // i-default
[maxLen]byte{'i', '-', 'e', 'n', 'o', 'c', 'h', 'i', 'a', 'n'}: -4, // i-enochian
[maxLen]byte{'i', '-', 'm', 'i', 'n', 'g', 'o'}: -5, // i-mingo
[maxLen]byte{'z', 'h', '-', 'm', 'i', 'n'}: -6, // zh-min
// CLDR-specific tag.
[maxLen]byte{'r', 'o', 'o', 't'}: 0, // root
[maxLen]byte{'e', 'n', '-', 'u', 's', '-', 'p', 'o', 's', 'i', 'x'}: -7, // en_US_POSIX"
}
altTagIndex = [...]uint8{0, 17, 31, 45, 61, 74, 86, 102}
altTags = "xtg-x-cel-gaulishen-GB-oxendicten-x-i-defaultund-x-i-enochiansee-x-i-mingonan-x-zh-minen-US-u-va-posix"
)
func grandfathered(s [maxAltTaglen]byte) (t Tag, ok bool) {
if v, ok := grandfatheredMap[s]; ok {
if v < 0 {
return Make(altTags[altTagIndex[-v-1]:altTagIndex[-v]]), true
}
t.lang = langID(v)
return t, true
}
return t, false
}

1648
vendor/golang.org/x/text/language/maketables.go generated vendored
View file

File diff suppressed because it is too large Load diff

841
vendor/golang.org/x/text/language/match.go generated vendored
View file

@ -1,841 +0,0 @@
// 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.
package language
import "errors"
// Matcher is the interface that wraps the Match method.
//
// Match returns the best match for any of the given tags, along with
// a unique index associated with the returned tag and a confidence
// score.
type Matcher interface {
Match(t ...Tag) (tag Tag, index int, c Confidence)
}
// Comprehends reports the confidence score for a speaker of a given language
// to being able to comprehend the written form of an alternative language.
func Comprehends(speaker, alternative Tag) Confidence {
_, _, c := NewMatcher([]Tag{alternative}).Match(speaker)
return c
}
// NewMatcher returns a Matcher that matches an ordered list of preferred tags
// against a list of supported tags based on written intelligibility, closeness
// of dialect, equivalence of subtags and various other rules. It is initialized
// with the list of supported tags. The first element is used as the default
// value in case no match is found.
//
// Its Match method matches the first of the given Tags to reach a certain
// confidence threshold. The tags passed to Match should therefore be specified
// in order of preference. Extensions are ignored for matching.
//
// The index returned by the Match method corresponds to the index of the
// matched tag in t, but is augmented with the Unicode extension ('u')of the
// corresponding preferred tag. This allows user locale options to be passed
// transparently.
func NewMatcher(t []Tag) Matcher {
return newMatcher(t)
}
func (m *matcher) Match(want ...Tag) (t Tag, index int, c Confidence) {
match, w, c := m.getBest(want...)
if match == nil {
t = m.default_.tag
} else {
t, index = match.tag, match.index
}
// Copy options from the user-provided tag into the result tag. This is hard
// to do after the fact, so we do it here.
// TODO: consider also adding in variants that are compatible with the
// matched language.
// TODO: Add back region if it is non-ambiguous? Or create another tag to
// preserve the region?
if u, ok := w.Extension('u'); ok {
t, _ = Raw.Compose(t, u)
}
return t, index, c
}
type scriptRegionFlags uint8
const (
isList = 1 << iota
scriptInFrom
regionInFrom
)
func (t *Tag) setUndefinedLang(id langID) {
if t.lang == 0 {
t.lang = id
}
}
func (t *Tag) setUndefinedScript(id scriptID) {
if t.script == 0 {
t.script = id
}
}
func (t *Tag) setUndefinedRegion(id regionID) {
if t.region == 0 || t.region.contains(id) {
t.region = id
}
}
// ErrMissingLikelyTagsData indicates no information was available
// to compute likely values of missing tags.
var ErrMissingLikelyTagsData = errors.New("missing likely tags data")
// addLikelySubtags sets subtags to their most likely value, given the locale.
// In most cases this means setting fields for unknown values, but in some
// cases it may alter a value. It returns a ErrMissingLikelyTagsData error
// if the given locale cannot be expanded.
func (t Tag) addLikelySubtags() (Tag, error) {
id, err := addTags(t)
if err != nil {
return t, err
} else if id.equalTags(t) {
return t, nil
}
id.remakeString()
return id, nil
}
// specializeRegion attempts to specialize a group region.
func specializeRegion(t *Tag) bool {
if i := regionInclusion[t.region]; i < nRegionGroups {
x := likelyRegionGroup[i]
if langID(x.lang) == t.lang && scriptID(x.script) == t.script {
t.region = regionID(x.region)
}
return true
}
return false
}
func addTags(t Tag) (Tag, error) {
// We leave private use identifiers alone.
if t.private() {
return t, nil
}
if t.script != 0 && t.region != 0 {
if t.lang != 0 {
// already fully specified
specializeRegion(&t)
return t, nil
}
// Search matches for und-script-region. Note that for these cases
// region will never be a group so there is no need to check for this.
list := likelyRegion[t.region : t.region+1]
if x := list[0]; x.flags&isList != 0 {
list = likelyRegionList[x.lang : x.lang+uint16(x.script)]
}
for _, x := range list {
// Deviating from the spec. See match_test.go for details.
if scriptID(x.script) == t.script {
t.setUndefinedLang(langID(x.lang))
return t, nil
}
}
}
if t.lang != 0 {
// Search matches for lang-script and lang-region, where lang != und.
if t.lang < langNoIndexOffset {
x := likelyLang[t.lang]
if x.flags&isList != 0 {
list := likelyLangList[x.region : x.region+uint16(x.script)]
if t.script != 0 {
for _, x := range list {
if scriptID(x.script) == t.script && x.flags&scriptInFrom != 0 {
t.setUndefinedRegion(regionID(x.region))
return t, nil
}
}
} else if t.region != 0 {
count := 0
goodScript := true
tt := t
for _, x := range list {
// We visit all entries for which the script was not
// defined, including the ones where the region was not
// defined. This allows for proper disambiguation within
// regions.
if x.flags&scriptInFrom == 0 && t.region.contains(regionID(x.region)) {
tt.region = regionID(x.region)
tt.setUndefinedScript(scriptID(x.script))
goodScript = goodScript && tt.script == scriptID(x.script)
count++
}
}
if count == 1 {
return tt, nil
}
// Even if we fail to find a unique Region, we might have
// an unambiguous script.
if goodScript {
t.script = tt.script
}
}
}
}
} else {
// Search matches for und-script.
if t.script != 0 {
x := likelyScript[t.script]
if x.region != 0 {
t.setUndefinedRegion(regionID(x.region))
t.setUndefinedLang(langID(x.lang))
return t, nil
}
}
// Search matches for und-region. If und-script-region exists, it would
// have been found earlier.
if t.region != 0 {
if i := regionInclusion[t.region]; i < nRegionGroups {
x := likelyRegionGroup[i]
if x.region != 0 {
t.setUndefinedLang(langID(x.lang))
t.setUndefinedScript(scriptID(x.script))
t.region = regionID(x.region)
}
} else {
x := likelyRegion[t.region]
if x.flags&isList != 0 {
x = likelyRegionList[x.lang]
}
if x.script != 0 && x.flags != scriptInFrom {
t.setUndefinedLang(langID(x.lang))
t.setUndefinedScript(scriptID(x.script))
return t, nil
}
}
}
}
// Search matches for lang.
if t.lang < langNoIndexOffset {
x := likelyLang[t.lang]
if x.flags&isList != 0 {
x = likelyLangList[x.region]
}
if x.region != 0 {
t.setUndefinedScript(scriptID(x.script))
t.setUndefinedRegion(regionID(x.region))
}
specializeRegion(&t)
if t.lang == 0 {
t.lang = _en // default language
}
return t, nil
}
return t, ErrMissingLikelyTagsData
}
func (t *Tag) setTagsFrom(id Tag) {
t.lang = id.lang
t.script = id.script
t.region = id.region
}
// minimize removes the region or script subtags from t such that
// t.addLikelySubtags() == t.minimize().addLikelySubtags().
func (t Tag) minimize() (Tag, error) {
t, err := minimizeTags(t)
if err != nil {
return t, err
}
t.remakeString()
return t, nil
}
// minimizeTags mimics the behavior of the ICU 51 C implementation.
func minimizeTags(t Tag) (Tag, error) {
if t.equalTags(und) {
return t, nil
}
max, err := addTags(t)
if err != nil {
return t, err
}
for _, id := range [...]Tag{
{lang: t.lang},
{lang: t.lang, region: t.region},
{lang: t.lang, script: t.script},
} {
if x, err := addTags(id); err == nil && max.equalTags(x) {
t.setTagsFrom(id)
break
}
}
return t, nil
}
// Tag Matching
// CLDR defines an algorithm for finding the best match between two sets of language
// tags. The basic algorithm defines how to score a possible match and then find
// the match with the best score
// (see http://www.unicode.org/reports/tr35/#LanguageMatching).
// Using scoring has several disadvantages. The scoring obfuscates the importance of
// the various factors considered, making the algorithm harder to understand. Using
// scoring also requires the full score to be computed for each pair of tags.
//
// We will use a different algorithm which aims to have the following properties:
// - clarity on the precedence of the various selection factors, and
// - improved performance by allowing early termination of a comparison.
//
// Matching algorithm (overview)
// Input:
// - supported: a set of supported tags
// - default: the default tag to return in case there is no match
// - desired: list of desired tags, ordered by preference, starting with
// the most-preferred.
//
// Algorithm:
// 1) Set the best match to the lowest confidence level
// 2) For each tag in "desired":
// a) For each tag in "supported":
// 1) compute the match between the two tags.
// 2) if the match is better than the previous best match, replace it
// with the new match. (see next section)
// b) if the current best match is above a certain threshold, return this
// match without proceeding to the next tag in "desired". [See Note 1]
// 3) If the best match so far is below a certain threshold, return "default".
//
// Ranking:
// We use two phases to determine whether one pair of tags are a better match
// than another pair of tags. First, we determine a rough confidence level. If the
// levels are different, the one with the highest confidence wins.
// Second, if the rough confidence levels are identical, we use a set of tie-breaker
// rules.
//
// The confidence level of matching a pair of tags is determined by finding the
// lowest confidence level of any matches of the corresponding subtags (the
// result is deemed as good as its weakest link).
// We define the following levels:
// Exact - An exact match of a subtag, before adding likely subtags.
// MaxExact - An exact match of a subtag, after adding likely subtags.
// [See Note 2].
// High - High level of mutual intelligibility between different subtag
// variants.
// Low - Low level of mutual intelligibility between different subtag
// variants.
// No - No mutual intelligibility.
//
// The following levels can occur for each type of subtag:
// Base: Exact, MaxExact, High, Low, No
// Script: Exact, MaxExact [see Note 3], Low, No
// Region: Exact, MaxExact, High
// Variant: Exact, High
// Private: Exact, No
//
// Any result with a confidence level of Low or higher is deemed a possible match.
// Once a desired tag matches any of the supported tags with a level of MaxExact
// or higher, the next desired tag is not considered (see Step 2.b).
// Note that CLDR provides languageMatching data that defines close equivalence
// classes for base languages, scripts and regions.
//
// Tie-breaking
// If we get the same confidence level for two matches, we apply a sequence of
// tie-breaking rules. The first that succeeds defines the result. The rules are
// applied in the following order.
// 1) Original language was defined and was identical.
// 2) Original region was defined and was identical.
// 3) Distance between two maximized regions was the smallest.
// 4) Original script was defined and was identical.
// 5) Distance from want tag to have tag using the parent relation [see Note 5.]
// If there is still no winner after these rules are applied, the first match
// found wins.
//
// Notes:
// [1] Note that even if we may not have a perfect match, if a match is above a
// certain threshold, it is considered a better match than any other match
// to a tag later in the list of preferred language tags.
// [2] In practice, as matching of Exact is done in a separate phase from
// matching the other levels, we reuse the Exact level to mean MaxExact in
// the second phase. As a consequence, we only need the levels defined by
// the Confidence type. The MaxExact confidence level is mapped to High in
// the public API.
// [3] We do not differentiate between maximized script values that were derived
// from suppressScript versus most likely tag data. We determined that in
// ranking the two, one ranks just after the other. Moreover, the two cannot
// occur concurrently. As a consequence, they are identical for practical
// purposes.
// [4] In case of deprecated, macro-equivalents and legacy mappings, we assign
// the MaxExact level to allow iw vs he to still be a closer match than
// en-AU vs en-US, for example.
// [5] In CLDR a locale inherits fields that are unspecified for this locale
// from its parent. Therefore, if a locale is a parent of another locale,
// it is a strong measure for closeness, especially when no other tie
// breaker rule applies. One could also argue it is inconsistent, for
// example, when pt-AO matches pt (which CLDR equates with pt-BR), even
// though its parent is pt-PT according to the inheritance rules.
//
// Implementation Details:
// There are several performance considerations worth pointing out. Most notably,
// we preprocess as much as possible (within reason) at the time of creation of a
// matcher. This includes:
// - creating a per-language map, which includes data for the raw base language
// and its canonicalized variant (if applicable),
// - expanding entries for the equivalence classes defined in CLDR's
// languageMatch data.
// The per-language map ensures that typically only a very small number of tags
// need to be considered. The pre-expansion of canonicalized subtags and
// equivalence classes reduces the amount of map lookups that need to be done at
// runtime.
// matcher keeps a set of supported language tags, indexed by language.
type matcher struct {
default_ *haveTag
index map[langID]*matchHeader
passSettings bool
}
// matchHeader has the lists of tags for exact matches and matches based on
// maximized and canonicalized tags for a given language.
type matchHeader struct {
exact []*haveTag
max []*haveTag
}
// haveTag holds a supported Tag and its maximized script and region. The maximized
// or canonicalized language is not stored as it is not needed during matching.
type haveTag struct {
tag Tag
// index of this tag in the original list of supported tags.
index int
// conf is the maximum confidence that can result from matching this haveTag.
// When conf < Exact this means it was inserted after applying a CLDR equivalence rule.
conf Confidence
// Maximized region and script.
maxRegion regionID
maxScript scriptID
// altScript may be checked as an alternative match to maxScript. If altScript
// matches, the confidence level for this match is Low. Theoretically there
// could be multiple alternative scripts. This does not occur in practice.
altScript scriptID
// nextMax is the index of the next haveTag with the same maximized tags.
nextMax uint16
}
func makeHaveTag(tag Tag, index int) (haveTag, langID) {
max := tag
if tag.lang != 0 {
max, _ = max.canonicalize(All)
max, _ = addTags(max)
max.remakeString()
}
return haveTag{tag, index, Exact, max.region, max.script, altScript(max.lang, max.script), 0}, max.lang
}
// altScript returns an alternative script that may match the given script with
// a low confidence. At the moment, the langMatch data allows for at most one
// script to map to another and we rely on this to keep the code simple.
func altScript(l langID, s scriptID) scriptID {
for _, alt := range matchScript {
if (alt.lang == 0 || langID(alt.lang) == l) && scriptID(alt.have) == s {
return scriptID(alt.want)
}
}
return 0
}
// addIfNew adds a haveTag to the list of tags only if it is a unique tag.
// Tags that have the same maximized values are linked by index.
func (h *matchHeader) addIfNew(n haveTag, exact bool) {
// Don't add new exact matches.
for _, v := range h.exact {
if v.tag.equalsRest(n.tag) {
return
}
}
if exact {
h.exact = append(h.exact, &n)
}
// Allow duplicate maximized tags, but create a linked list to allow quickly
// comparing the equivalents and bail out.
for i, v := range h.max {
if v.maxScript == n.maxScript &&
v.maxRegion == n.maxRegion &&
v.tag.variantOrPrivateTagStr() == n.tag.variantOrPrivateTagStr() {
for h.max[i].nextMax != 0 {
i = int(h.max[i].nextMax)
}
h.max[i].nextMax = uint16(len(h.max))
break
}
}
h.max = append(h.max, &n)
}
// header returns the matchHeader for the given language. It creates one if
// it doesn't already exist.
func (m *matcher) header(l langID) *matchHeader {
if h := m.index[l]; h != nil {
return h
}
h := &matchHeader{}
m.index[l] = h
return h
}
// newMatcher builds an index for the given supported tags and returns it as
// a matcher. It also expands the index by considering various equivalence classes
// for a given tag.
func newMatcher(supported []Tag) *matcher {
m := &matcher{
index: make(map[langID]*matchHeader),
}
if len(supported) == 0 {
m.default_ = &haveTag{}
return m
}
// Add supported languages to the index. Add exact matches first to give
// them precedence.
for i, tag := range supported {
pair, _ := makeHaveTag(tag, i)
m.header(tag.lang).addIfNew(pair, true)
}
m.default_ = m.header(supported[0].lang).exact[0]
for i, tag := range supported {
pair, max := makeHaveTag(tag, i)
if max != tag.lang {
m.header(max).addIfNew(pair, false)
}
}
// update is used to add indexes in the map for equivalent languages.
// If force is true, the update will also apply to derived entries. To
// avoid applying a "transitive closure", use false.
update := func(want, have uint16, conf Confidence, force bool) {
if hh := m.index[langID(have)]; hh != nil {
if !force && len(hh.exact) == 0 {
return
}
hw := m.header(langID(want))
for _, ht := range hh.max {
v := *ht
if conf < v.conf {
v.conf = conf
}
v.nextMax = 0 // this value needs to be recomputed
if v.altScript != 0 {
v.altScript = altScript(langID(want), v.maxScript)
}
hw.addIfNew(v, conf == Exact && len(hh.exact) > 0)
}
}
}
// Add entries for languages with mutual intelligibility as defined by CLDR's
// languageMatch data.
for _, ml := range matchLang {
update(ml.want, ml.have, Confidence(ml.conf), false)
if !ml.oneway {
update(ml.have, ml.want, Confidence(ml.conf), false)
}
}
// Add entries for possible canonicalizations. This is an optimization to
// ensure that only one map lookup needs to be done at runtime per desired tag.
// First we match deprecated equivalents. If they are perfect equivalents
// (their canonicalization simply substitutes a different language code, but
// nothing else), the match confidence is Exact, otherwise it is High.
for i, lm := range langAliasMap {
if lm.from == _sh {
continue
}
// If deprecated codes match and there is no fiddling with the script or
// or region, we consider it an exact match.
conf := Exact
if langAliasTypes[i] != langMacro {
if !isExactEquivalent(langID(lm.from)) {
conf = High
}
update(lm.to, lm.from, conf, true)
}
update(lm.from, lm.to, conf, true)
}
return m
}
// getBest gets the best matching tag in m for any of the given tags, taking into
// account the order of preference of the given tags.
func (m *matcher) getBest(want ...Tag) (got *haveTag, orig Tag, c Confidence) {
best := bestMatch{}
for _, w := range want {
var max Tag
// Check for exact match first.
h := m.index[w.lang]
if w.lang != 0 {
// Base language is defined.
if h == nil {
continue
}
for i := range h.exact {
have := h.exact[i]
if have.tag.equalsRest(w) {
return have, w, Exact
}
}
max, _ = w.canonicalize(Legacy | Deprecated)
max, _ = addTags(max)
} else {
// Base language is not defined.
if h != nil {
for i := range h.exact {
have := h.exact[i]
if have.tag.equalsRest(w) {
return have, w, Exact
}
}
}
if w.script == 0 && w.region == 0 {
// We skip all tags matching und for approximate matching, including
// private tags.
continue
}
max, _ = addTags(w)
if h = m.index[max.lang]; h == nil {
continue
}
}
// Check for match based on maximized tag.
for i := range h.max {
have := h.max[i]
best.update(have, w, max.script, max.region)
if best.conf == Exact {
for have.nextMax != 0 {
have = h.max[have.nextMax]
best.update(have, w, max.script, max.region)
}
return best.have, best.want, High
}
}
}
if best.conf <= No {
if len(want) != 0 {
return nil, want[0], No
}
return nil, Tag{}, No
}
return best.have, best.want, best.conf
}
// bestMatch accumulates the best match so far.
type bestMatch struct {
have *haveTag
want Tag
conf Confidence
// Cached results from applying tie-breaking rules.
origLang bool
origReg bool
regDist uint8
origScript bool
parentDist uint8 // 255 if have is not an ancestor of want tag.
}
// update updates the existing best match if the new pair is considered to be a
// better match.
// To determine if the given pair is a better match, it first computes the rough
// confidence level. If this surpasses the current match, it will replace it and
// update the tie-breaker rule cache. If there is a tie, it proceeds with applying
// a series of tie-breaker rules. If there is no conclusive winner after applying
// the tie-breaker rules, it leaves the current match as the preferred match.
func (m *bestMatch) update(have *haveTag, tag Tag, maxScript scriptID, maxRegion regionID) {
// Bail if the maximum attainable confidence is below that of the current best match.
c := have.conf
if c < m.conf {
return
}
if have.maxScript != maxScript {
// There is usually very little comprehension between different scripts.
// In a few cases there may still be Low comprehension. This possibility is
// pre-computed and stored in have.altScript.
if Low < m.conf || have.altScript != maxScript {
return
}
c = Low
} else if have.maxRegion != maxRegion {
// There is usually a small difference between languages across regions.
// We use the region distance (below) to disambiguate between equal matches.
if High < c {
c = High
}
}
// We store the results of the computations of the tie-breaker rules along
// with the best match. There is no need to do the checks once we determine
// we have a winner, but we do still need to do the tie-breaker computations.
// We use "beaten" to keep track if we still need to do the checks.
beaten := false // true if the new pair defeats the current one.
if c != m.conf {
if c < m.conf {
return
}
beaten = true
}
// Tie-breaker rules:
// We prefer if the pre-maximized language was specified and identical.
origLang := have.tag.lang == tag.lang && tag.lang != 0
if !beaten && m.origLang != origLang {
if m.origLang {
return
}
beaten = true
}
// We prefer if the pre-maximized region was specified and identical.
origReg := have.tag.region == tag.region && tag.region != 0
if !beaten && m.origReg != origReg {
if m.origReg {
return
}
beaten = true
}
// Next we prefer smaller distances between regions, as defined by regionDist.
regDist := regionDist(have.maxRegion, maxRegion, tag.lang)
if !beaten && m.regDist != regDist {
if regDist > m.regDist {
return
}
beaten = true
}
// Next we prefer if the pre-maximized script was specified and identical.
origScript := have.tag.script == tag.script && tag.script != 0
if !beaten && m.origScript != origScript {
if m.origScript {
return
}
beaten = true
}
// Finally we prefer tags which have a closer parent relationship.
parentDist := parentDistance(have.tag.region, tag)
if !beaten && m.parentDist != parentDist {
if parentDist > m.parentDist {
return
}
beaten = true
}
// Update m to the newly found best match.
if beaten {
m.have = have
m.want = tag
m.conf = c
m.origLang = origLang
m.origReg = origReg
m.origScript = origScript
m.regDist = regDist
m.parentDist = parentDist
}
}
// parentDistance returns the number of times Parent must be called before the
// regions match. It is assumed that it has already been checked that lang and
// script are identical. If haveRegion does not occur in the ancestor chain of
// tag, it returns 255.
func parentDistance(haveRegion regionID, tag Tag) uint8 {
p := tag.Parent()
d := uint8(1)
for haveRegion != p.region {
if p.region == 0 {
return 255
}
p = p.Parent()
d++
}
return d
}
// regionDist wraps regionDistance with some exceptions to the algorithmic distance.
func regionDist(a, b regionID, lang langID) uint8 {
if lang == _en {
// Two variants of non-US English are close to each other, regardless of distance.
if a != _US && b != _US {
return 2
}
}
return uint8(regionDistance(a, b))
}
// regionDistance computes the distance between two regions based on the
// distance in the graph of region containments as defined in CLDR. It iterates
// over increasingly inclusive sets of groups, represented as bit vectors, until
// the source bit vector has bits in common with the destination vector.
func regionDistance(a, b regionID) int {
if a == b {
return 0
}
p, q := regionInclusion[a], regionInclusion[b]
if p < nRegionGroups {
p, q = q, p
}
set := regionInclusionBits
if q < nRegionGroups && set[p]&(1<<q) != 0 {
return 1
}
d := 2
for goal := set[q]; set[p]&goal == 0; p = regionInclusionNext[p] {
d++
}
return d
}
func (t Tag) variants() string {
if t.pVariant == 0 {
return ""
}
return t.str[t.pVariant:t.pExt]
}
// variantOrPrivateTagStr returns variants or private use tags.
func (t Tag) variantOrPrivateTagStr() string {
if t.pExt > 0 {
return t.str[t.pVariant:t.pExt]
}
return t.str[t.pVariant:]
}
// equalsRest compares everything except the language.
func (a Tag) equalsRest(b Tag) bool {
// TODO: don't include extensions in this comparison. To do this efficiently,
// though, we should handle private tags separately.
return a.script == b.script && a.region == b.region && a.variantOrPrivateTagStr() == b.variantOrPrivateTagStr()
}
// isExactEquivalent returns true if canonicalizing the language will not alter
// the script or region of a tag.
func isExactEquivalent(l langID) bool {
for _, o := range notEquivalent {
if o == l {
return false
}
}
return true
}
var notEquivalent []langID
func init() {
// Create a list of all languages for which canonicalization may alter the
// script or region.
for _, lm := range langAliasMap {
tag := Tag{lang: langID(lm.from)}
if tag, _ = tag.canonicalize(All); tag.script != 0 || tag.region != 0 {
notEquivalent = append(notEquivalent, langID(lm.from))
}
}
}

859
vendor/golang.org/x/text/language/parse.go generated vendored
View file

@ -1,859 +0,0 @@
// 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.
package language
import (
"bytes"
"errors"
"fmt"
"sort"
"strconv"
"strings"
"golang.org/x/text/internal/tag"
)
// isAlpha returns true if the byte is not a digit.
// b must be an ASCII letter or digit.
func isAlpha(b byte) bool {
return b > '9'
}
// isAlphaNum returns true if the string contains only ASCII letters or digits.
func isAlphaNum(s []byte) bool {
for _, c := range s {
if !('a' <= c && c <= 'z' || 'A' <= c && c <= 'Z' || '0' <= c && c <= '9') {
return false
}
}
return true
}
// errSyntax is returned by any of the parsing functions when the
// input is not well-formed, according to BCP 47.
// TODO: return the position at which the syntax error occurred?
var errSyntax = errors.New("language: tag is not well-formed")
// ValueError is returned by any of the parsing functions when the
// input is well-formed but the respective subtag is not recognized
// as a valid value.
type ValueError struct {
v [8]byte
}
func mkErrInvalid(s []byte) error {
var e ValueError
copy(e.v[:], s)
return e
}
func (e ValueError) tag() []byte {
n := bytes.IndexByte(e.v[:], 0)
if n == -1 {
n = 8
}
return e.v[:n]
}
// Error implements the error interface.
func (e ValueError) Error() string {
return fmt.Sprintf("language: subtag %q is well-formed but unknown", e.tag())
}
// Subtag returns the subtag for which the error occurred.
func (e ValueError) Subtag() string {
return string(e.tag())
}
// scanner is used to scan BCP 47 tokens, which are separated by _ or -.
type scanner struct {
b []byte
bytes [max99thPercentileSize]byte
token []byte
start int // start position of the current token
end int // end position of the current token
next int // next point for scan
err error
done bool
}
func makeScannerString(s string) scanner {
scan := scanner{}
if len(s) <= len(scan.bytes) {
scan.b = scan.bytes[:copy(scan.bytes[:], s)]
} else {
scan.b = []byte(s)
}
scan.init()
return scan
}
// makeScanner returns a scanner using b as the input buffer.
// b is not copied and may be modified by the scanner routines.
func makeScanner(b []byte) scanner {
scan := scanner{b: b}
scan.init()
return scan
}
func (s *scanner) init() {
for i, c := range s.b {
if c == '_' {
s.b[i] = '-'
}
}
s.scan()
}
// restToLower converts the string between start and end to lower case.
func (s *scanner) toLower(start, end int) {
for i := start; i < end; i++ {
c := s.b[i]
if 'A' <= c && c <= 'Z' {
s.b[i] += 'a' - 'A'
}
}
}
func (s *scanner) setError(e error) {
if s.err == nil || (e == errSyntax && s.err != errSyntax) {
s.err = e
}
}
// resizeRange shrinks or grows the array at position oldStart such that
// a new string of size newSize can fit between oldStart and oldEnd.
// Sets the scan point to after the resized range.
func (s *scanner) resizeRange(oldStart, oldEnd, newSize int) {
s.start = oldStart
if end := oldStart + newSize; end != oldEnd {
diff := end - oldEnd
if end < cap(s.b) {
b := make([]byte, len(s.b)+diff)
copy(b, s.b[:oldStart])
copy(b[end:], s.b[oldEnd:])
s.b = b
} else {
s.b = append(s.b[end:], s.b[oldEnd:]...)
}
s.next = end + (s.next - s.end)
s.end = end
}
}
// replace replaces the current token with repl.
func (s *scanner) replace(repl string) {
s.resizeRange(s.start, s.end, len(repl))
copy(s.b[s.start:], repl)
}
// gobble removes the current token from the input.
// Caller must call scan after calling gobble.
func (s *scanner) gobble(e error) {
s.setError(e)
if s.start == 0 {
s.b = s.b[:+copy(s.b, s.b[s.next:])]
s.end = 0
} else {
s.b = s.b[:s.start-1+copy(s.b[s.start-1:], s.b[s.end:])]
s.end = s.start - 1
}
s.next = s.start
}
// deleteRange removes the given range from s.b before the current token.
func (s *scanner) deleteRange(start, end int) {
s.setError(errSyntax)
s.b = s.b[:start+copy(s.b[start:], s.b[end:])]
diff := end - start
s.next -= diff
s.start -= diff
s.end -= diff
}
// scan parses the next token of a BCP 47 string. Tokens that are larger
// than 8 characters or include non-alphanumeric characters result in an error
// and are gobbled and removed from the output.
// It returns the end position of the last token consumed.
func (s *scanner) scan() (end int) {
end = s.end
s.token = nil
for s.start = s.next; s.next < len(s.b); {
i := bytes.IndexByte(s.b[s.next:], '-')
if i == -1 {
s.end = len(s.b)
s.next = len(s.b)
i = s.end - s.start
} else {
s.end = s.next + i
s.next = s.end + 1
}
token := s.b[s.start:s.end]
if i < 1 || i > 8 || !isAlphaNum(token) {
s.gobble(errSyntax)
continue
}
s.token = token
return end
}
if n := len(s.b); n > 0 && s.b[n-1] == '-' {
s.setError(errSyntax)
s.b = s.b[:len(s.b)-1]
}
s.done = true
return end
}
// acceptMinSize parses multiple tokens of the given size or greater.
// It returns the end position of the last token consumed.
func (s *scanner) acceptMinSize(min int) (end int) {
end = s.end
s.scan()
for ; len(s.token) >= min; s.scan() {
end = s.end
}
return end
}
// Parse parses the given BCP 47 string and returns a valid Tag. If parsing
// failed it returns an error and any part of the tag that could be parsed.
// If parsing succeeded but an unknown value was found, it returns
// ValueError. The Tag returned in this case is just stripped of the unknown
// value. All other values are preserved. It accepts tags in the BCP 47 format
// and extensions to this standard defined in
// http://www.unicode.org/reports/tr35/#Unicode_Language_and_Locale_Identifiers.
// The resulting tag is canonicalized using the default canonicalization type.
func Parse(s string) (t Tag, err error) {
return Default.Parse(s)
}
// Parse parses the given BCP 47 string and returns a valid Tag. If parsing
// failed it returns an error and any part of the tag that could be parsed.
// If parsing succeeded but an unknown value was found, it returns
// ValueError. The Tag returned in this case is just stripped of the unknown
// value. All other values are preserved. It accepts tags in the BCP 47 format
// and extensions to this standard defined in
// http://www.unicode.org/reports/tr35/#Unicode_Language_and_Locale_Identifiers.
// The resulting tag is canonicalized using the the canonicalization type c.
func (c CanonType) Parse(s string) (t Tag, err error) {
// TODO: consider supporting old-style locale key-value pairs.
if s == "" {
return und, errSyntax
}
if len(s) <= maxAltTaglen {
b := [maxAltTaglen]byte{}
for i, c := range s {
// Generating invalid UTF-8 is okay as it won't match.
if 'A' <= c && c <= 'Z' {
c += 'a' - 'A'
} else if c == '_' {
c = '-'
}
b[i] = byte(c)
}
if t, ok := grandfathered(b); ok {
return t, nil
}
}
scan := makeScannerString(s)
t, err = parse(&scan, s)
t, changed := t.canonicalize(c)
if changed {
t.remakeString()
}
return t, err
}
func parse(scan *scanner, s string) (t Tag, err error) {
t = und
var end int
if n := len(scan.token); n <= 1 {
scan.toLower(0, len(scan.b))
if n == 0 || scan.token[0] != 'x' {
return t, errSyntax
}
end = parseExtensions(scan)
} else if n >= 4 {
return und, errSyntax
} else { // the usual case
t, end = parseTag(scan)
if n := len(scan.token); n == 1 {
t.pExt = uint16(end)
end = parseExtensions(scan)
} else if end < len(scan.b) {
scan.setError(errSyntax)
scan.b = scan.b[:end]
}
}
if int(t.pVariant) < len(scan.b) {
if end < len(s) {
s = s[:end]
}
if len(s) > 0 && tag.Compare(s, scan.b) == 0 {
t.str = s
} else {
t.str = string(scan.b)
}
} else {
t.pVariant, t.pExt = 0, 0
}
return t, scan.err
}
// parseTag parses language, script, region and variants.
// It returns a Tag and the end position in the input that was parsed.
func parseTag(scan *scanner) (t Tag, end int) {
var e error
// TODO: set an error if an unknown lang, script or region is encountered.
t.lang, e = getLangID(scan.token)
scan.setError(e)
scan.replace(t.lang.String())
langStart := scan.start
end = scan.scan()
for len(scan.token) == 3 && isAlpha(scan.token[0]) {
// From http://tools.ietf.org/html/bcp47, <lang>-<extlang> tags are equivalent
// to a tag of the form <extlang>.
lang, e := getLangID(scan.token)
if lang != 0 {
t.lang = lang
copy(scan.b[langStart:], lang.String())
scan.b[langStart+3] = '-'
scan.start = langStart + 4
}
scan.gobble(e)
end = scan.scan()
}
if len(scan.token) == 4 && isAlpha(scan.token[0]) {
t.script, e = getScriptID(script, scan.token)
if t.script == 0 {
scan.gobble(e)
}
end = scan.scan()
}
if n := len(scan.token); n >= 2 && n <= 3 {
t.region, e = getRegionID(scan.token)
if t.region == 0 {
scan.gobble(e)
} else {
scan.replace(t.region.String())
}
end = scan.scan()
}
scan.toLower(scan.start, len(scan.b))
t.pVariant = byte(end)
end = parseVariants(scan, end, t)
t.pExt = uint16(end)
return t, end
}
var separator = []byte{'-'}
// parseVariants scans tokens as long as each token is a valid variant string.
// Duplicate variants are removed.
func parseVariants(scan *scanner, end int, t Tag) int {
start := scan.start
varIDBuf := [4]uint8{}
variantBuf := [4][]byte{}
varID := varIDBuf[:0]
variant := variantBuf[:0]
last := -1
needSort := false
for ; len(scan.token) >= 4; scan.scan() {
// TODO: measure the impact of needing this conversion and redesign
// the data structure if there is an issue.
v, ok := variantIndex[string(scan.token)]
if !ok {
// unknown variant
// TODO: allow user-defined variants?
scan.gobble(mkErrInvalid(scan.token))
continue
}
varID = append(varID, v)
variant = append(variant, scan.token)
if !needSort {
if last < int(v) {
last = int(v)
} else {
needSort = true
// There is no legal combinations of more than 7 variants
// (and this is by no means a useful sequence).
const maxVariants = 8
if len(varID) > maxVariants {
break
}
}
}
end = scan.end
}
if needSort {
sort.Sort(variantsSort{varID, variant})
k, l := 0, -1
for i, v := range varID {
w := int(v)
if l == w {
// Remove duplicates.
continue
}
varID[k] = varID[i]
variant[k] = variant[i]
k++
l = w
}
if str := bytes.Join(variant[:k], separator); len(str) == 0 {
end = start - 1
} else {
scan.resizeRange(start, end, len(str))
copy(scan.b[scan.start:], str)
end = scan.end
}
}
return end
}
type variantsSort struct {
i []uint8
v [][]byte
}
func (s variantsSort) Len() int {
return len(s.i)
}
func (s variantsSort) Swap(i, j int) {
s.i[i], s.i[j] = s.i[j], s.i[i]
s.v[i], s.v[j] = s.v[j], s.v[i]
}
func (s variantsSort) Less(i, j int) bool {
return s.i[i] < s.i[j]
}
type bytesSort [][]byte
func (b bytesSort) Len() int {
return len(b)
}
func (b bytesSort) Swap(i, j int) {
b[i], b[j] = b[j], b[i]
}
func (b bytesSort) Less(i, j int) bool {
return bytes.Compare(b[i], b[j]) == -1
}
// parseExtensions parses and normalizes the extensions in the buffer.
// It returns the last position of scan.b that is part of any extension.
// It also trims scan.b to remove excess parts accordingly.
func parseExtensions(scan *scanner) int {
start := scan.start
exts := [][]byte{}
private := []byte{}
end := scan.end
for len(scan.token) == 1 {
extStart := scan.start
ext := scan.token[0]
end = parseExtension(scan)
extension := scan.b[extStart:end]
if len(extension) < 3 || (ext != 'x' && len(extension) < 4) {
scan.setError(errSyntax)
end = extStart
continue
} else if start == extStart && (ext == 'x' || scan.start == len(scan.b)) {
scan.b = scan.b[:end]
return end
} else if ext == 'x' {
private = extension
break
}
exts = append(exts, extension)
}
sort.Sort(bytesSort(exts))
if len(private) > 0 {
exts = append(exts, private)
}
scan.b = scan.b[:start]
if len(exts) > 0 {
scan.b = append(scan.b, bytes.Join(exts, separator)...)
} else if start > 0 {
// Strip trailing '-'.
scan.b = scan.b[:start-1]
}
return end
}
// parseExtension parses a single extension and returns the position of
// the extension end.
func parseExtension(scan *scanner) int {
start, end := scan.start, scan.end
switch scan.token[0] {
case 'u':
attrStart := end
scan.scan()
for last := []byte{}; len(scan.token) > 2; scan.scan() {
if bytes.Compare(scan.token, last) != -1 {
// Attributes are unsorted. Start over from scratch.
p := attrStart + 1
scan.next = p
attrs := [][]byte{}
for scan.scan(); len(scan.token) > 2; scan.scan() {
attrs = append(attrs, scan.token)
end = scan.end
}
sort.Sort(bytesSort(attrs))
copy(scan.b[p:], bytes.Join(attrs, separator))
break
}
last = scan.token
end = scan.end
}
var last, key []byte
for attrEnd := end; len(scan.token) == 2; last = key {
key = scan.token
keyEnd := scan.end
end = scan.acceptMinSize(3)
// TODO: check key value validity
if keyEnd == end || bytes.Compare(key, last) != 1 {
// We have an invalid key or the keys are not sorted.
// Start scanning keys from scratch and reorder.
p := attrEnd + 1
scan.next = p
keys := [][]byte{}
for scan.scan(); len(scan.token) == 2; {
keyStart, keyEnd := scan.start, scan.end
end = scan.acceptMinSize(3)
if keyEnd != end {
keys = append(keys, scan.b[keyStart:end])
} else {
scan.setError(errSyntax)
end = keyStart
}
}
sort.Sort(bytesSort(keys))
reordered := bytes.Join(keys, separator)
if e := p + len(reordered); e < end {
scan.deleteRange(e, end)
end = e
}
copy(scan.b[p:], bytes.Join(keys, separator))
break
}
}
case 't':
scan.scan()
if n := len(scan.token); n >= 2 && n <= 3 && isAlpha(scan.token[1]) {
_, end = parseTag(scan)
scan.toLower(start, end)
}
for len(scan.token) == 2 && !isAlpha(scan.token[1]) {
end = scan.acceptMinSize(3)
}
case 'x':
end = scan.acceptMinSize(1)
default:
end = scan.acceptMinSize(2)
}
return end
}
// Compose creates a Tag from individual parts, which may be of type Tag, Base,
// Script, Region, Variant, []Variant, Extension, []Extension or error. If a
// Base, Script or Region or slice of type Variant or Extension is passed more
// than once, the latter will overwrite the former. Variants and Extensions are
// accumulated, but if two extensions of the same type are passed, the latter
// will replace the former. A Tag overwrites all former values and typically
// only makes sense as the first argument. The resulting tag is returned after
// canonicalizing using the Default CanonType. If one or more errors are
// encountered, one of the errors is returned.
func Compose(part ...interface{}) (t Tag, err error) {
return Default.Compose(part...)
}
// Compose creates a Tag from individual parts, which may be of type Tag, Base,
// Script, Region, Variant, []Variant, Extension, []Extension or error. If a
// Base, Script or Region or slice of type Variant or Extension is passed more
// than once, the latter will overwrite the former. Variants and Extensions are
// accumulated, but if two extensions of the same type are passed, the latter
// will replace the former. A Tag overwrites all former values and typically
// only makes sense as the first argument. The resulting tag is returned after
// canonicalizing using CanonType c. If one or more errors are encountered,
// one of the errors is returned.
func (c CanonType) Compose(part ...interface{}) (t Tag, err error) {
var b builder
if err = b.update(part...); err != nil {
return und, err
}
t, _ = b.tag.canonicalize(c)
if len(b.ext) > 0 || len(b.variant) > 0 {
sort.Sort(sortVariant(b.variant))
sort.Strings(b.ext)
if b.private != "" {
b.ext = append(b.ext, b.private)
}
n := maxCoreSize + tokenLen(b.variant...) + tokenLen(b.ext...)
buf := make([]byte, n)
p := t.genCoreBytes(buf)
t.pVariant = byte(p)
p += appendTokens(buf[p:], b.variant...)
t.pExt = uint16(p)
p += appendTokens(buf[p:], b.ext...)
t.str = string(buf[:p])
} else if b.private != "" {
t.str = b.private
t.remakeString()
}
return
}
type builder struct {
tag Tag
private string // the x extension
ext []string
variant []string
err error
}
func (b *builder) addExt(e string) {
if e == "" {
} else if e[0] == 'x' {
b.private = e
} else {
b.ext = append(b.ext, e)
}
}
var errInvalidArgument = errors.New("invalid Extension or Variant")
func (b *builder) update(part ...interface{}) (err error) {
replace := func(l *[]string, s string, eq func(a, b string) bool) bool {
if s == "" {
b.err = errInvalidArgument
return true
}
for i, v := range *l {
if eq(v, s) {
(*l)[i] = s
return true
}
}
return false
}
for _, x := range part {
switch v := x.(type) {
case Tag:
b.tag.lang = v.lang
b.tag.region = v.region
b.tag.script = v.script
if v.str != "" {
b.variant = nil
for x, s := "", v.str[v.pVariant:v.pExt]; s != ""; {
x, s = nextToken(s)
b.variant = append(b.variant, x)
}
b.ext, b.private = nil, ""
for i, e := int(v.pExt), ""; i < len(v.str); {
i, e = getExtension(v.str, i)
b.addExt(e)
}
}
case Base:
b.tag.lang = v.langID
case Script:
b.tag.script = v.scriptID
case Region:
b.tag.region = v.regionID
case Variant:
if !replace(&b.variant, v.variant, func(a, b string) bool { return a == b }) {
b.variant = append(b.variant, v.variant)
}
case Extension:
if !replace(&b.ext, v.s, func(a, b string) bool { return a[0] == b[0] }) {
b.addExt(v.s)
}
case []Variant:
b.variant = nil
for _, x := range v {
b.update(x)
}
case []Extension:
b.ext, b.private = nil, ""
for _, e := range v {
b.update(e)
}
// TODO: support parsing of raw strings based on morphology or just extensions?
case error:
err = v
}
}
return
}
func tokenLen(token ...string) (n int) {
for _, t := range token {
n += len(t) + 1
}
return
}
func appendTokens(b []byte, token ...string) int {
p := 0
for _, t := range token {
b[p] = '-'
copy(b[p+1:], t)
p += 1 + len(t)
}
return p
}
type sortVariant []string
func (s sortVariant) Len() int {
return len(s)
}
func (s sortVariant) Swap(i, j int) {
s[j], s[i] = s[i], s[j]
}
func (s sortVariant) Less(i, j int) bool {
return variantIndex[s[i]] < variantIndex[s[j]]
}
func findExt(list []string, x byte) int {
for i, e := range list {
if e[0] == x {
return i
}
}
return -1
}
// getExtension returns the name, body and end position of the extension.
func getExtension(s string, p int) (end int, ext string) {
if s[p] == '-' {
p++
}
if s[p] == 'x' {
return len(s), s[p:]
}
end = nextExtension(s, p)
return end, s[p:end]
}
// nextExtension finds the next extension within the string, searching
// for the -<char>- pattern from position p.
// In the fast majority of cases, language tags will have at most
// one extension and extensions tend to be small.
func nextExtension(s string, p int) int {
for n := len(s) - 3; p < n; {
if s[p] == '-' {
if s[p+2] == '-' {
return p
}
p += 3
} else {
p++
}
}
return len(s)
}
var errInvalidWeight = errors.New("ParseAcceptLanguage: invalid weight")
// ParseAcceptLanguage parses the contents of a Accept-Language header as
// defined in http://www.ietf.org/rfc/rfc2616.txt and returns a list of Tags and
// a list of corresponding quality weights. It is more permissive than RFC 2616
// and may return non-nil slices even if the input is not valid.
// The Tags will be sorted by highest weight first and then by first occurrence.
// Tags with a weight of zero will be dropped. An error will be returned if the
// input could not be parsed.
func ParseAcceptLanguage(s string) (tag []Tag, q []float32, err error) {
var entry string
for s != "" {
if entry, s = split(s, ','); entry == "" {
continue
}
entry, weight := split(entry, ';')
// Scan the language.
t, err := Parse(entry)
if err != nil {
id, ok := acceptFallback[entry]
if !ok {
return nil, nil, err
}
t = Tag{lang: id}
}
// Scan the optional weight.
w := 1.0
if weight != "" {
weight = consume(weight, 'q')
weight = consume(weight, '=')
// consume returns the empty string when a token could not be
// consumed, resulting in an error for ParseFloat.
if w, err = strconv.ParseFloat(weight, 32); err != nil {
return nil, nil, errInvalidWeight
}
// Drop tags with a quality weight of 0.
if w <= 0 {
continue
}
}
tag = append(tag, t)
q = append(q, float32(w))
}
sortStable(&tagSort{tag, q})
return tag, q, nil
}
// consume removes a leading token c from s and returns the result or the empty
// string if there is no such token.
func consume(s string, c byte) string {
if s == "" || s[0] != c {
return ""
}
return strings.TrimSpace(s[1:])
}
func split(s string, c byte) (head, tail string) {
if i := strings.IndexByte(s, c); i >= 0 {
return strings.TrimSpace(s[:i]), strings.TrimSpace(s[i+1:])
}
return strings.TrimSpace(s), ""
}
// Add hack mapping to deal with a small number of cases that that occur
// in Accept-Language (with reasonable frequency).
var acceptFallback = map[string]langID{
"english": _en,
"deutsch": _de,
"italian": _it,
"french": _fr,
"*": _mul, // defined in the spec to match all languages.
}
type tagSort struct {
tag []Tag
q []float32
}
func (s *tagSort) Len() int {
return len(s.q)
}
func (s *tagSort) Less(i, j int) bool {
return s.q[i] > s.q[j]
}
func (s *tagSort) Swap(i, j int) {
s.tag[i], s.tag[j] = s.tag[j], s.tag[i]
s.q[i], s.q[j] = s.q[j], s.q[i]
}

3547
vendor/golang.org/x/text/language/tables.go generated vendored
View file

File diff suppressed because it is too large Load diff

143
vendor/golang.org/x/text/language/tags.go generated vendored
View file

@ -1,143 +0,0 @@
// 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.
package language
// TODO: Various sets of commonly use tags and regions.
// MustParse is like Parse, but panics if the given BCP 47 tag cannot be parsed.
// It simplifies safe initialization of Tag values.
func MustParse(s string) Tag {
t, err := Parse(s)
if err != nil {
panic(err)
}
return t
}
// MustParse is like Parse, but panics if the given BCP 47 tag cannot be parsed.
// It simplifies safe initialization of Tag values.
func (c CanonType) MustParse(s string) Tag {
t, err := c.Parse(s)
if err != nil {
panic(err)
}
return t
}
// MustParseBase is like ParseBase, but panics if the given base cannot be parsed.
// It simplifies safe initialization of Base values.
func MustParseBase(s string) Base {
b, err := ParseBase(s)
if err != nil {
panic(err)
}
return b
}
// MustParseScript is like ParseScript, but panics if the given script cannot be
// parsed. It simplifies safe initialization of Script values.
func MustParseScript(s string) Script {
scr, err := ParseScript(s)
if err != nil {
panic(err)
}
return scr
}
// MustParseRegion is like ParseRegion, but panics if the given region cannot be
// parsed. It simplifies safe initialization of Region values.
func MustParseRegion(s string) Region {
r, err := ParseRegion(s)
if err != nil {
panic(err)
}
return r
}
var (
und = Tag{}
Und Tag = Tag{}
Afrikaans Tag = Tag{lang: _af} // af
Amharic Tag = Tag{lang: _am} // am
Arabic Tag = Tag{lang: _ar} // ar
ModernStandardArabic Tag = Tag{lang: _ar, region: _001} // ar-001
Azerbaijani Tag = Tag{lang: _az} // az
Bulgarian Tag = Tag{lang: _bg} // bg
Bengali Tag = Tag{lang: _bn} // bn
Catalan Tag = Tag{lang: _ca} // ca
Czech Tag = Tag{lang: _cs} // cs
Danish Tag = Tag{lang: _da} // da
German Tag = Tag{lang: _de} // de
Greek Tag = Tag{lang: _el} // el
English Tag = Tag{lang: _en} // en
AmericanEnglish Tag = Tag{lang: _en, region: _US} // en-US
BritishEnglish Tag = Tag{lang: _en, region: _GB} // en-GB
Spanish Tag = Tag{lang: _es} // es
EuropeanSpanish Tag = Tag{lang: _es, region: _ES} // es-ES
LatinAmericanSpanish Tag = Tag{lang: _es, region: _419} // es-419
Estonian Tag = Tag{lang: _et} // et
Persian Tag = Tag{lang: _fa} // fa
Finnish Tag = Tag{lang: _fi} // fi
Filipino Tag = Tag{lang: _fil} // fil
French Tag = Tag{lang: _fr} // fr
CanadianFrench Tag = Tag{lang: _fr, region: _CA} // fr-CA
Gujarati Tag = Tag{lang: _gu} // gu
Hebrew Tag = Tag{lang: _he} // he
Hindi Tag = Tag{lang: _hi} // hi
Croatian Tag = Tag{lang: _hr} // hr
Hungarian Tag = Tag{lang: _hu} // hu
Armenian Tag = Tag{lang: _hy} // hy
Indonesian Tag = Tag{lang: _id} // id
Icelandic Tag = Tag{lang: _is} // is
Italian Tag = Tag{lang: _it} // it
Japanese Tag = Tag{lang: _ja} // ja
Georgian Tag = Tag{lang: _ka} // ka
Kazakh Tag = Tag{lang: _kk} // kk
Khmer Tag = Tag{lang: _km} // km
Kannada Tag = Tag{lang: _kn} // kn
Korean Tag = Tag{lang: _ko} // ko
Kirghiz Tag = Tag{lang: _ky} // ky
Lao Tag = Tag{lang: _lo} // lo
Lithuanian Tag = Tag{lang: _lt} // lt
Latvian Tag = Tag{lang: _lv} // lv
Macedonian Tag = Tag{lang: _mk} // mk
Malayalam Tag = Tag{lang: _ml} // ml
Mongolian Tag = Tag{lang: _mn} // mn
Marathi Tag = Tag{lang: _mr} // mr
Malay Tag = Tag{lang: _ms} // ms
Burmese Tag = Tag{lang: _my} // my
Nepali Tag = Tag{lang: _ne} // ne
Dutch Tag = Tag{lang: _nl} // nl
Norwegian Tag = Tag{lang: _no} // no
Punjabi Tag = Tag{lang: _pa} // pa
Polish Tag = Tag{lang: _pl} // pl
Portuguese Tag = Tag{lang: _pt} // pt
BrazilianPortuguese Tag = Tag{lang: _pt, region: _BR} // pt-BR
EuropeanPortuguese Tag = Tag{lang: _pt, region: _PT} // pt-PT
Romanian Tag = Tag{lang: _ro} // ro
Russian Tag = Tag{lang: _ru} // ru
Sinhala Tag = Tag{lang: _si} // si
Slovak Tag = Tag{lang: _sk} // sk
Slovenian Tag = Tag{lang: _sl} // sl
Albanian Tag = Tag{lang: _sq} // sq
Serbian Tag = Tag{lang: _sr} // sr
SerbianLatin Tag = Tag{lang: _sr, script: _Latn} // sr-Latn
Swedish Tag = Tag{lang: _sv} // sv
Swahili Tag = Tag{lang: _sw} // sw
Tamil Tag = Tag{lang: _ta} // ta
Telugu Tag = Tag{lang: _te} // te
Thai Tag = Tag{lang: _th} // th
Turkish Tag = Tag{lang: _tr} // tr
Ukrainian Tag = Tag{lang: _uk} // uk
Urdu Tag = Tag{lang: _ur} // ur
Uzbek Tag = Tag{lang: _uz} // uz
Vietnamese Tag = Tag{lang: _vi} // vi
Chinese Tag = Tag{lang: _zh} // zh
SimplifiedChinese Tag = Tag{lang: _zh, script: _Hans} // zh-Hans
TraditionalChinese Tag = Tag{lang: _zh, script: _Hant} // zh-Hant
Zulu Tag = Tag{lang: _zu} // zu
)

187
vendor/golang.org/x/text/runes/cond.go generated vendored
View file

@ -1,187 +0,0 @@
// 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.
package runes
import (
"unicode/utf8"
"golang.org/x/text/transform"
)
// Note: below we pass invalid UTF-8 to the tIn and tNotIn transformers as is.
// This is done for various reasons:
// - To retain the semantics of the Nop transformer: if input is passed to a Nop
// one would expect it to be unchanged.
// - It would be very expensive to pass a converted RuneError to a transformer:
// a transformer might need more source bytes after RuneError, meaning that
// the only way to pass it safely is to create a new buffer and manage the
// intermingling of RuneErrors and normal input.
// - Many transformers leave ill-formed UTF-8 as is, so this is not
// inconsistent. Generally ill-formed UTF-8 is only replaced if it is a
// logical consequence of the operation (as for Map) or if it otherwise would
// pose security concerns (as for Remove).
// - An alternative would be to return an error on ill-formed UTF-8, but this
// would be inconsistent with other operations.
// If returns a transformer that applies tIn to consecutive runes for which
// s.Contains(r) and tNotIn to consecutive runes for which !s.Contains(r). Reset
// is called on tIn and tNotIn at the start of each run. A Nop transformer will
// substitute a nil value passed to tIn or tNotIn. Invalid UTF-8 is translated
// to RuneError to determine which transformer to apply, but is passed as is to
// the respective transformer.
func If(s Set, tIn, tNotIn transform.Transformer) Transformer {
if tIn == nil && tNotIn == nil {
return Transformer{transform.Nop}
}
if tIn == nil {
tIn = transform.Nop
}
if tNotIn == nil {
tNotIn = transform.Nop
}
sIn, ok := tIn.(transform.SpanningTransformer)
if !ok {
sIn = dummySpan{tIn}
}
sNotIn, ok := tNotIn.(transform.SpanningTransformer)
if !ok {
sNotIn = dummySpan{tNotIn}
}
a := &cond{
tIn: sIn,
tNotIn: sNotIn,
f: s.Contains,
}
a.Reset()
return Transformer{a}
}
type dummySpan struct{ transform.Transformer }
func (d dummySpan) Span(src []byte, atEOF bool) (n int, err error) {
return 0, transform.ErrEndOfSpan
}
type cond struct {
tIn, tNotIn transform.SpanningTransformer
f func(rune) bool
check func(rune) bool // current check to perform
t transform.SpanningTransformer // current transformer to use
}
// Reset implements transform.Transformer.
func (t *cond) Reset() {
t.check = t.is
t.t = t.tIn
t.t.Reset() // notIn will be reset on first usage.
}
func (t *cond) is(r rune) bool {
if t.f(r) {
return true
}
t.check = t.isNot
t.t = t.tNotIn
t.tNotIn.Reset()
return false
}
func (t *cond) isNot(r rune) bool {
if !t.f(r) {
return true
}
t.check = t.is
t.t = t.tIn
t.tIn.Reset()
return false
}
// This implementation of Span doesn't help all too much, but it needs to be
// there to satisfy this package's Transformer interface.
// TODO: there are certainly room for improvements, though. For example, if
// t.t == transform.Nop (which will a common occurrence) it will save a bundle
// to special-case that loop.
func (t *cond) Span(src []byte, atEOF bool) (n int, err error) {
p := 0
for n < len(src) && err == nil {
// Don't process too much at a time as the Spanner that will be
// called on this block may terminate early.
const maxChunk = 4096
max := len(src)
if v := n + maxChunk; v < max {
max = v
}
atEnd := false
size := 0
current := t.t
for ; p < max; p += size {
r := rune(src[p])
if r < utf8.RuneSelf {
size = 1
} else if r, size = utf8.DecodeRune(src[p:]); size == 1 {
if !atEOF && !utf8.FullRune(src[p:]) {
err = transform.ErrShortSrc
break
}
}
if !t.check(r) {
// The next rune will be the start of a new run.
atEnd = true
break
}
}
n2, err2 := current.Span(src[n:p], atEnd || (atEOF && p == len(src)))
n += n2
if err2 != nil {
return n, err2
}
// At this point either err != nil or t.check will pass for the rune at p.
p = n + size
}
return n, err
}
func (t *cond) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
p := 0
for nSrc < len(src) && err == nil {
// Don't process too much at a time, as the work might be wasted if the
// destination buffer isn't large enough to hold the result or a
// transform returns an error early.
const maxChunk = 4096
max := len(src)
if n := nSrc + maxChunk; n < len(src) {
max = n
}
atEnd := false
size := 0
current := t.t
for ; p < max; p += size {
r := rune(src[p])
if r < utf8.RuneSelf {
size = 1
} else if r, size = utf8.DecodeRune(src[p:]); size == 1 {
if !atEOF && !utf8.FullRune(src[p:]) {
err = transform.ErrShortSrc
break
}
}
if !t.check(r) {
// The next rune will be the start of a new run.
atEnd = true
break
}
}
nDst2, nSrc2, err2 := current.Transform(dst[nDst:], src[nSrc:p], atEnd || (atEOF && p == len(src)))
nDst += nDst2
nSrc += nSrc2
if err2 != nil {
return nDst, nSrc, err2
}
// At this point either err != nil or t.check will pass for the rune at p.
p = nSrc + size
}
return nDst, nSrc, err
}

355
vendor/golang.org/x/text/runes/runes.go generated vendored
View file

@ -1,355 +0,0 @@
// Copyright 2014 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 runes provide transforms for UTF-8 encoded text.
package runes // import "golang.org/x/text/runes"
import (
"unicode"
"unicode/utf8"
"golang.org/x/text/transform"
)
// A Set is a collection of runes.
type Set interface {
// Contains returns true if r is contained in the set.
Contains(r rune) bool
}
type setFunc func(rune) bool
func (s setFunc) Contains(r rune) bool {
return s(r)
}
// Note: using funcs here instead of wrapping types result in cleaner
// documentation and a smaller API.
// In creates a Set with a Contains method that returns true for all runes in
// the given RangeTable.
func In(rt *unicode.RangeTable) Set {
return setFunc(func(r rune) bool { return unicode.Is(rt, r) })
}
// In creates a Set with a Contains method that returns true for all runes not
// in the given RangeTable.
func NotIn(rt *unicode.RangeTable) Set {
return setFunc(func(r rune) bool { return !unicode.Is(rt, r) })
}
// Predicate creates a Set with a Contains method that returns f(r).
func Predicate(f func(rune) bool) Set {
return setFunc(f)
}
// Transformer implements the transform.Transformer interface.
type Transformer struct {
t transform.SpanningTransformer
}
func (t Transformer) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
return t.t.Transform(dst, src, atEOF)
}
func (t Transformer) Span(b []byte, atEOF bool) (n int, err error) {
return t.t.Span(b, atEOF)
}
func (t Transformer) Reset() { t.t.Reset() }
// Bytes returns a new byte slice with the result of converting b using t. It
// calls Reset on t. It returns nil if any error was found. This can only happen
// if an error-producing Transformer is passed to If.
func (t Transformer) Bytes(b []byte) []byte {
b, _, err := transform.Bytes(t, b)
if err != nil {
return nil
}
return b
}
// String returns a string with the result of converting s using t. It calls
// Reset on t. It returns the empty string if any error was found. This can only
// happen if an error-producing Transformer is passed to If.
func (t Transformer) String(s string) string {
s, _, err := transform.String(t, s)
if err != nil {
return ""
}
return s
}
// TODO:
// - Copy: copying strings and bytes in whole-rune units.
// - Validation (maybe)
// - Well-formed-ness (maybe)
const runeErrorString = string(utf8.RuneError)
// Remove returns a Transformer that removes runes r for which s.Contains(r).
// Illegal input bytes are replaced by RuneError before being passed to f.
func Remove(s Set) Transformer {
if f, ok := s.(setFunc); ok {
// This little trick cuts the running time of BenchmarkRemove for sets
// created by Predicate roughly in half.
// TODO: special-case RangeTables as well.
return Transformer{remove(f)}
}
return Transformer{remove(s.Contains)}
}
// TODO: remove transform.RemoveFunc.
type remove func(r rune) bool
func (remove) Reset() {}
// Span implements transform.Spanner.
func (t remove) Span(src []byte, atEOF bool) (n int, err error) {
for r, size := rune(0), 0; n < len(src); {
if r = rune(src[n]); r < utf8.RuneSelf {
size = 1
} else if r, size = utf8.DecodeRune(src[n:]); size == 1 {
// Invalid rune.
if !atEOF && !utf8.FullRune(src[n:]) {
err = transform.ErrShortSrc
} else {
err = transform.ErrEndOfSpan
}
break
}
if t(r) {
err = transform.ErrEndOfSpan
break
}
n += size
}
return
}
// Transform implements transform.Transformer.
func (t remove) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
for r, size := rune(0), 0; nSrc < len(src); {
if r = rune(src[nSrc]); r < utf8.RuneSelf {
size = 1
} else if r, size = utf8.DecodeRune(src[nSrc:]); size == 1 {
// Invalid rune.
if !atEOF && !utf8.FullRune(src[nSrc:]) {
err = transform.ErrShortSrc
break
}
// We replace illegal bytes with RuneError. Not doing so might
// otherwise turn a sequence of invalid UTF-8 into valid UTF-8.
// The resulting byte sequence may subsequently contain runes
// for which t(r) is true that were passed unnoticed.
if !t(utf8.RuneError) {
if nDst+3 > len(dst) {
err = transform.ErrShortDst
break
}
dst[nDst+0] = runeErrorString[0]
dst[nDst+1] = runeErrorString[1]
dst[nDst+2] = runeErrorString[2]
nDst += 3
}
nSrc++
continue
}
if t(r) {
nSrc += size
continue
}
if nDst+size > len(dst) {
err = transform.ErrShortDst
break
}
for i := 0; i < size; i++ {
dst[nDst] = src[nSrc]
nDst++
nSrc++
}
}
return
}
// Map returns a Transformer that maps the runes in the input using the given
// mapping. Illegal bytes in the input are converted to utf8.RuneError before
// being passed to the mapping func.
func Map(mapping func(rune) rune) Transformer {
return Transformer{mapper(mapping)}
}
type mapper func(rune) rune
func (mapper) Reset() {}
// Span implements transform.Spanner.
func (t mapper) Span(src []byte, atEOF bool) (n int, err error) {
for r, size := rune(0), 0; n < len(src); n += size {
if r = rune(src[n]); r < utf8.RuneSelf {
size = 1
} else if r, size = utf8.DecodeRune(src[n:]); size == 1 {
// Invalid rune.
if !atEOF && !utf8.FullRune(src[n:]) {
err = transform.ErrShortSrc
} else {
err = transform.ErrEndOfSpan
}
break
}
if t(r) != r {
err = transform.ErrEndOfSpan
break
}
}
return n, err
}
// Transform implements transform.Transformer.
func (t mapper) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
var replacement rune
var b [utf8.UTFMax]byte
for r, size := rune(0), 0; nSrc < len(src); {
if r = rune(src[nSrc]); r < utf8.RuneSelf {
if replacement = t(r); replacement < utf8.RuneSelf {
if nDst == len(dst) {
err = transform.ErrShortDst
break
}
dst[nDst] = byte(replacement)
nDst++
nSrc++
continue
}
size = 1
} else if r, size = utf8.DecodeRune(src[nSrc:]); size == 1 {
// Invalid rune.
if !atEOF && !utf8.FullRune(src[nSrc:]) {
err = transform.ErrShortSrc
break
}
if replacement = t(utf8.RuneError); replacement == utf8.RuneError {
if nDst+3 > len(dst) {
err = transform.ErrShortDst
break
}
dst[nDst+0] = runeErrorString[0]
dst[nDst+1] = runeErrorString[1]
dst[nDst+2] = runeErrorString[2]
nDst += 3
nSrc++
continue
}
} else if replacement = t(r); replacement == r {
if nDst+size > len(dst) {
err = transform.ErrShortDst
break
}
for i := 0; i < size; i++ {
dst[nDst] = src[nSrc]
nDst++
nSrc++
}
continue
}
n := utf8.EncodeRune(b[:], replacement)
if nDst+n > len(dst) {
err = transform.ErrShortDst
break
}
for i := 0; i < n; i++ {
dst[nDst] = b[i]
nDst++
}
nSrc += size
}
return
}
// ReplaceIllFormed returns a transformer that replaces all input bytes that are
// not part of a well-formed UTF-8 code sequence with utf8.RuneError.
func ReplaceIllFormed() Transformer {
return Transformer{&replaceIllFormed{}}
}
type replaceIllFormed struct{ transform.NopResetter }
func (t replaceIllFormed) Span(src []byte, atEOF bool) (n int, err error) {
for n < len(src) {
// ASCII fast path.
if src[n] < utf8.RuneSelf {
n++
continue
}
r, size := utf8.DecodeRune(src[n:])
// Look for a valid non-ASCII rune.
if r != utf8.RuneError || size != 1 {
n += size
continue
}
// Look for short source data.
if !atEOF && !utf8.FullRune(src[n:]) {
err = transform.ErrShortSrc
break
}
// We have an invalid rune.
err = transform.ErrEndOfSpan
break
}
return n, err
}
func (t replaceIllFormed) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
for nSrc < len(src) {
// ASCII fast path.
if r := src[nSrc]; r < utf8.RuneSelf {
if nDst == len(dst) {
err = transform.ErrShortDst
break
}
dst[nDst] = r
nDst++
nSrc++
continue
}
// Look for a valid non-ASCII rune.
if _, size := utf8.DecodeRune(src[nSrc:]); size != 1 {
if size != copy(dst[nDst:], src[nSrc:nSrc+size]) {
err = transform.ErrShortDst
break
}
nDst += size
nSrc += size
continue
}
// Look for short source data.
if !atEOF && !utf8.FullRune(src[nSrc:]) {
err = transform.ErrShortSrc
break
}
// We have an invalid rune.
if nDst+3 > len(dst) {
err = transform.ErrShortDst
break
}
dst[nDst+0] = runeErrorString[0]
dst[nDst+1] = runeErrorString[1]
dst[nDst+2] = runeErrorString[2]
nDst += 3
nSrc++
}
return nDst, nSrc, err
}

36
vendor/golang.org/x/text/secure/precis/class.go generated vendored
View file

@ -1,36 +0,0 @@
// 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.
package precis
import (
"unicode/utf8"
)
// TODO: Add contextual character rules from Appendix A of RFC5892.
// A class is a set of characters that match certain derived properties. The
// PRECIS framework defines two classes: The Freeform class and the Identifier
// class. The freeform class should be used for profiles where expressiveness is
// prioritized over safety such as nicknames or passwords. The identifier class
// should be used for profiles where safety is the first priority such as
// addressable network labels and usernames.
type class struct {
validFrom property
}
// Contains satisfies the runes.Set interface and returns whether the given rune
// is a member of the class.
func (c class) Contains(r rune) bool {
b := make([]byte, 4)
n := utf8.EncodeRune(b, r)
trieval, _ := dpTrie.lookup(b[:n])
return c.validFrom <= property(trieval)
}
var (
identifier = &class{validFrom: pValid}
freeform = &class{validFrom: idDisOrFreePVal}
)

139
vendor/golang.org/x/text/secure/precis/context.go generated vendored
View file

@ -1,139 +0,0 @@
// Copyright 2016 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 precis
import "errors"
// This file contains tables and code related to context rules.
type catBitmap uint16
const (
// These bits, once set depending on the current value, are never unset.
bJapanese catBitmap = 1 << iota
bArabicIndicDigit
bExtendedArabicIndicDigit
// These bits are set on each iteration depending on the current value.
bJoinStart
bJoinMid
bJoinEnd
bVirama
bLatinSmallL
bGreek
bHebrew
// These bits indicated which of the permanent bits need to be set at the
// end of the checks.
bMustHaveJapn
permanent = bJapanese | bArabicIndicDigit | bExtendedArabicIndicDigit | bMustHaveJapn
)
const finalShift = 10
var errContext = errors.New("precis: contextual rule violated")
func init() {
// Programmatically set these required bits as, manually setting them seems
// too error prone.
for i, ct := range categoryTransitions {
categoryTransitions[i].keep |= permanent
categoryTransitions[i].accept |= ct.term
}
}
var categoryTransitions = []struct {
keep catBitmap // mask selecting which bits to keep from the previous state
set catBitmap // mask for which bits to set for this transition
// These bitmaps are used for rules that require lookahead.
// term&accept == term must be true, which is enforced programmatically.
term catBitmap // bits accepted as termination condition
accept catBitmap // bits that pass, but not sufficient as termination
// The rule function cannot take a *context as an argument, as it would
// cause the context to escape, adding significant overhead.
rule func(beforeBits catBitmap) (doLookahead bool, err error)
}{
joiningL: {set: bJoinStart},
joiningD: {set: bJoinStart | bJoinEnd},
joiningT: {keep: bJoinStart, set: bJoinMid},
joiningR: {set: bJoinEnd},
viramaModifier: {set: bVirama},
viramaJoinT: {set: bVirama | bJoinMid},
latinSmallL: {set: bLatinSmallL},
greek: {set: bGreek},
greekJoinT: {set: bGreek | bJoinMid},
hebrew: {set: bHebrew},
hebrewJoinT: {set: bHebrew | bJoinMid},
japanese: {set: bJapanese},
katakanaMiddleDot: {set: bMustHaveJapn},
zeroWidthNonJoiner: {
term: bJoinEnd,
accept: bJoinMid,
rule: func(before catBitmap) (doLookAhead bool, err error) {
if before&bVirama != 0 {
return false, nil
}
if before&bJoinStart == 0 {
return false, errContext
}
return true, nil
},
},
zeroWidthJoiner: {
rule: func(before catBitmap) (doLookAhead bool, err error) {
if before&bVirama == 0 {
err = errContext
}
return false, err
},
},
middleDot: {
term: bLatinSmallL,
rule: func(before catBitmap) (doLookAhead bool, err error) {
if before&bLatinSmallL == 0 {
return false, errContext
}
return true, nil
},
},
greekLowerNumeralSign: {
set: bGreek,
term: bGreek,
rule: func(before catBitmap) (doLookAhead bool, err error) {
return true, nil
},
},
hebrewPreceding: {
set: bHebrew,
rule: func(before catBitmap) (doLookAhead bool, err error) {
if before&bHebrew == 0 {
err = errContext
}
return false, err
},
},
arabicIndicDigit: {
set: bArabicIndicDigit,
rule: func(before catBitmap) (doLookAhead bool, err error) {
if before&bExtendedArabicIndicDigit != 0 {
err = errContext
}
return false, err
},
},
extendedArabicIndicDigit: {
set: bExtendedArabicIndicDigit,
rule: func(before catBitmap) (doLookAhead bool, err error) {
if before&bArabicIndicDigit != 0 {
err = errContext
}
return false, err
},
},
}

14
vendor/golang.org/x/text/secure/precis/doc.go generated vendored
View file

@ -1,14 +0,0 @@
// 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.
// Package precis contains types and functions for the preparation,
// enforcement, and comparison of internationalized strings ("PRECIS") as
// defined in RFC 7564. It also contains several pre-defined profiles for
// passwords, nicknames, and usernames as defined in RFC 7613 and RFC 7700.
//
// BE ADVISED: This package is under construction and the API may change in
// backwards incompatible ways and without notice.
package precis // import "golang.org/x/text/secure/precis"
//go:generate go run gen.go gen_trieval.go

310
vendor/golang.org/x/text/secure/precis/gen.go generated vendored
View file

@ -1,310 +0,0 @@
// 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.
// Unicode table generator.
// Data read from the web.
// +build ignore
package main
import (
"flag"
"log"
"unicode"
"unicode/utf8"
"golang.org/x/text/internal/gen"
"golang.org/x/text/internal/triegen"
"golang.org/x/text/internal/ucd"
"golang.org/x/text/unicode/norm"
"golang.org/x/text/unicode/rangetable"
)
var outputFile = flag.String("output", "tables.go", "output file for generated tables; default tables.go")
var assigned, disallowedRunes *unicode.RangeTable
var runeCategory = map[rune]category{}
var overrides = map[category]category{
viramaModifier: viramaJoinT,
greek: greekJoinT,
hebrew: hebrewJoinT,
}
func setCategory(r rune, cat category) {
if c, ok := runeCategory[r]; ok {
if override, ok := overrides[c]; cat == joiningT && ok {
cat = override
} else {
log.Fatalf("%U: multiple categories for rune (%v and %v)", r, c, cat)
}
}
runeCategory[r] = cat
}
func init() {
if numCategories > 1<<propShift {
log.Fatalf("Number of categories is %d; may at most be %d", numCategories, 1<<propShift)
}
}
func main() {
gen.Init()
// Load data
runes := []rune{}
// PrecisIgnorableProperties: https://tools.ietf.org/html/rfc7564#section-9.13
ucd.Parse(gen.OpenUCDFile("DerivedCoreProperties.txt"), func(p *ucd.Parser) {
if p.String(1) == "Default_Ignorable_Code_Point" {
runes = append(runes, p.Rune(0))
}
})
ucd.Parse(gen.OpenUCDFile("PropList.txt"), func(p *ucd.Parser) {
switch p.String(1) {
case "Noncharacter_Code_Point":
runes = append(runes, p.Rune(0))
}
})
// OldHangulJamo: https://tools.ietf.org/html/rfc5892#section-2.9
ucd.Parse(gen.OpenUCDFile("HangulSyllableType.txt"), func(p *ucd.Parser) {
switch p.String(1) {
case "L", "V", "T":
runes = append(runes, p.Rune(0))
}
})
disallowedRunes = rangetable.New(runes...)
assigned = rangetable.Assigned(unicode.Version)
// Load category data.
runeCategory['l'] = latinSmallL
ucd.Parse(gen.OpenUCDFile("UnicodeData.txt"), func(p *ucd.Parser) {
const cccVirama = 9
if p.Int(ucd.CanonicalCombiningClass) == cccVirama {
setCategory(p.Rune(0), viramaModifier)
}
})
ucd.Parse(gen.OpenUCDFile("Scripts.txt"), func(p *ucd.Parser) {
switch p.String(1) {
case "Greek":
setCategory(p.Rune(0), greek)
case "Hebrew":
setCategory(p.Rune(0), hebrew)
case "Hiragana", "Katakana", "Han":
setCategory(p.Rune(0), japanese)
}
})
// Set the rule categories associated with exceptions. This overrides any
// previously set categories. The original categories are manually
// reintroduced in the categoryTransitions table.
for r, e := range exceptions {
if e.cat != 0 {
runeCategory[r] = e.cat
}
}
cat := map[string]category{
"L": joiningL,
"D": joiningD,
"T": joiningT,
"R": joiningR,
}
ucd.Parse(gen.OpenUCDFile("extracted/DerivedJoiningType.txt"), func(p *ucd.Parser) {
switch v := p.String(1); v {
case "L", "D", "T", "R":
setCategory(p.Rune(0), cat[v])
}
})
writeTables()
gen.Repackage("gen_trieval.go", "trieval.go", "precis")
}
type exception struct {
prop property
cat category
}
func init() {
// Programmatically add the Arabic and Indic digits to the exceptions map.
// See comment in the exceptions map below why these are marked disallowed.
for i := rune(0); i <= 9; i++ {
exceptions[0x0660+i] = exception{
prop: disallowed,
cat: arabicIndicDigit,
}
exceptions[0x06F0+i] = exception{
prop: disallowed,
cat: extendedArabicIndicDigit,
}
}
}
// The Exceptions class as defined in RFC 5892
// https://tools.ietf.org/html/rfc5892#section-2.6
var exceptions = map[rune]exception{
0x00DF: {prop: pValid},
0x03C2: {prop: pValid},
0x06FD: {prop: pValid},
0x06FE: {prop: pValid},
0x0F0B: {prop: pValid},
0x3007: {prop: pValid},
// ContextO|J rules are marked as disallowed, taking a "guilty until proven
// innocent" approach. The main reason for this is that the check for
// whether a context rule should be applied can be moved to the logic for
// handing disallowed runes, taken it off the common path. The exception to
// this rule is for katakanaMiddleDot, as the rule logic is handled without
// using a rule function.
// ContextJ (Join control)
0x200C: {prop: disallowed, cat: zeroWidthNonJoiner},
0x200D: {prop: disallowed, cat: zeroWidthJoiner},
// ContextO
0x00B7: {prop: disallowed, cat: middleDot},
0x0375: {prop: disallowed, cat: greekLowerNumeralSign},
0x05F3: {prop: disallowed, cat: hebrewPreceding}, // punctuation Geresh
0x05F4: {prop: disallowed, cat: hebrewPreceding}, // punctuation Gershayim
0x30FB: {prop: pValid, cat: katakanaMiddleDot},
// These are officially ContextO, but the implementation does not require
// special treatment of these, so we simply mark them as valid.
0x0660: {prop: pValid},
0x0661: {prop: pValid},
0x0662: {prop: pValid},
0x0663: {prop: pValid},
0x0664: {prop: pValid},
0x0665: {prop: pValid},
0x0666: {prop: pValid},
0x0667: {prop: pValid},
0x0668: {prop: pValid},
0x0669: {prop: pValid},
0x06F0: {prop: pValid},
0x06F1: {prop: pValid},
0x06F2: {prop: pValid},
0x06F3: {prop: pValid},
0x06F4: {prop: pValid},
0x06F5: {prop: pValid},
0x06F6: {prop: pValid},
0x06F7: {prop: pValid},
0x06F8: {prop: pValid},
0x06F9: {prop: pValid},
0x0640: {prop: disallowed},
0x07FA: {prop: disallowed},
0x302E: {prop: disallowed},
0x302F: {prop: disallowed},
0x3031: {prop: disallowed},
0x3032: {prop: disallowed},
0x3033: {prop: disallowed},
0x3034: {prop: disallowed},
0x3035: {prop: disallowed},
0x303B: {prop: disallowed},
}
// LetterDigits: https://tools.ietf.org/html/rfc5892#section-2.1
// r in {Ll, Lu, Lo, Nd, Lm, Mn, Mc}.
func isLetterDigits(r rune) bool {
return unicode.In(r,
unicode.Ll, unicode.Lu, unicode.Lm, unicode.Lo, // Letters
unicode.Mn, unicode.Mc, // Modifiers
unicode.Nd, // Digits
)
}
func isIdDisAndFreePVal(r rune) bool {
return unicode.In(r,
// OtherLetterDigits: https://tools.ietf.org/html/rfc7564#section-9.18
// r in in {Lt, Nl, No, Me}
unicode.Lt, unicode.Nl, unicode.No, // Other letters / numbers
unicode.Me, // Modifiers
// Spaces: https://tools.ietf.org/html/rfc7564#section-9.14
// r in in {Zs}
unicode.Zs,
// Symbols: https://tools.ietf.org/html/rfc7564#section-9.15
// r in {Sm, Sc, Sk, So}
unicode.Sm, unicode.Sc, unicode.Sk, unicode.So,
// Punctuation: https://tools.ietf.org/html/rfc7564#section-9.16
// r in {Pc, Pd, Ps, Pe, Pi, Pf, Po}
unicode.Pc, unicode.Pd, unicode.Ps, unicode.Pe,
unicode.Pi, unicode.Pf, unicode.Po,
)
}
// HasCompat: https://tools.ietf.org/html/rfc7564#section-9.17
func hasCompat(r rune) bool {
return !norm.NFKC.IsNormalString(string(r))
}
// From https://tools.ietf.org/html/rfc5892:
//
// If .cp. .in. Exceptions Then Exceptions(cp);
// Else If .cp. .in. BackwardCompatible Then BackwardCompatible(cp);
// Else If .cp. .in. Unassigned Then UNASSIGNED;
// Else If .cp. .in. ASCII7 Then PVALID;
// Else If .cp. .in. JoinControl Then CONTEXTJ;
// Else If .cp. .in. OldHangulJamo Then DISALLOWED;
// Else If .cp. .in. PrecisIgnorableProperties Then DISALLOWED;
// Else If .cp. .in. Controls Then DISALLOWED;
// Else If .cp. .in. HasCompat Then ID_DIS or FREE_PVAL;
// Else If .cp. .in. LetterDigits Then PVALID;
// Else If .cp. .in. OtherLetterDigits Then ID_DIS or FREE_PVAL;
// Else If .cp. .in. Spaces Then ID_DIS or FREE_PVAL;
// Else If .cp. .in. Symbols Then ID_DIS or FREE_PVAL;
// Else If .cp. .in. Punctuation Then ID_DIS or FREE_PVAL;
// Else DISALLOWED;
func writeTables() {
propTrie := triegen.NewTrie("derivedProperties")
w := gen.NewCodeWriter()
defer w.WriteGoFile(*outputFile, "precis")
gen.WriteUnicodeVersion(w)
// Iterate over all the runes...
for i := rune(0); i < unicode.MaxRune; i++ {
r := rune(i)
if !utf8.ValidRune(r) {
continue
}
e, ok := exceptions[i]
p := e.prop
switch {
case ok:
case !unicode.In(r, assigned):
p = unassigned
case r >= 0x0021 && r <= 0x007e: // Is ASCII 7
p = pValid
case unicode.In(r, disallowedRunes, unicode.Cc):
p = disallowed
case hasCompat(r):
p = idDisOrFreePVal
case isLetterDigits(r):
p = pValid
case isIdDisAndFreePVal(r):
p = idDisOrFreePVal
default:
p = disallowed
}
cat := runeCategory[r]
// Don't set category for runes that are disallowed.
if p == disallowed {
cat = exceptions[r].cat
}
propTrie.Insert(r, uint64(p)|uint64(cat))
}
sz, err := propTrie.Gen(w)
if err != nil {
log.Fatal(err)
}
w.Size += sz
}

68
vendor/golang.org/x/text/secure/precis/gen_trieval.go generated vendored
View file

@ -1,68 +0,0 @@
// 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
// entry is the entry of a trie table
// 7..6 property (unassigned, disallowed, maybe, valid)
// 5..0 category
type entry uint8
const (
propShift = 6
propMask = 0xc0
catMask = 0x3f
)
func (e entry) property() property { return property(e & propMask) }
func (e entry) category() category { return category(e & catMask) }
type property uint8
// The order of these constants matter. A Profile may consider runes to be
// allowed either from pValid or idDisOrFreePVal.
const (
unassigned property = iota << propShift
disallowed
idDisOrFreePVal // disallowed for Identifier, pValid for FreeForm
pValid
)
// compute permutations of all properties and specialCategories.
type category uint8
const (
other category = iota
// Special rune types
joiningL
joiningD
joiningT
joiningR
viramaModifier
viramaJoinT // Virama + JoiningT
latinSmallL // U+006c
greek
greekJoinT // Greek + JoiningT
hebrew
hebrewJoinT // Hebrew + JoiningT
japanese // hirigana, katakana, han
// Special rune types associated with contextual rules defined in
// https://tools.ietf.org/html/rfc5892#appendix-A.
// ContextO
zeroWidthNonJoiner // rule 1
zeroWidthJoiner // rule 2
// ContextJ
middleDot // rule 3
greekLowerNumeralSign // rule 4
hebrewPreceding // rule 5 and 6
katakanaMiddleDot // rule 7
arabicIndicDigit // rule 8
extendedArabicIndicDigit // rule 9
numCategories
)

70
vendor/golang.org/x/text/secure/precis/nickname.go generated vendored
View file

@ -1,70 +0,0 @@
// 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.
package precis
import (
"unicode"
"unicode/utf8"
"golang.org/x/text/transform"
)
type nickAdditionalMapping struct {
// TODO: This transformer needs to be stateless somehow…
notStart bool
prevSpace bool
}
func (t *nickAdditionalMapping) Reset() {
t.prevSpace = false
t.notStart = false
}
func (t *nickAdditionalMapping) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
// RFC 7700 §2.1. Rules
//
// 2. Additional Mapping Rule: The additional mapping rule consists of
// the following sub-rules.
//
// 1. Any instances of non-ASCII space MUST be mapped to ASCII
// space (U+0020); a non-ASCII space is any Unicode code point
// having a general category of "Zs", naturally with the
// exception of U+0020.
//
// 2. Any instances of the ASCII space character at the beginning
// or end of a nickname MUST be removed (e.g., "stpeter " is
// mapped to "stpeter").
//
// 3. Interior sequences of more than one ASCII space character
// MUST be mapped to a single ASCII space character (e.g.,
// "St Peter" is mapped to "St Peter").
for nSrc < len(src) {
r, size := utf8.DecodeRune(src[nSrc:])
if size == 0 { // Incomplete UTF-8 encoding
if !atEOF {
return nDst, nSrc, transform.ErrShortSrc
}
size = 1
}
if unicode.Is(unicode.Zs, r) {
t.prevSpace = true
} else {
if t.prevSpace && t.notStart {
dst[nDst] = ' '
nDst += 1
}
if size != copy(dst[nDst:], src[nSrc:nSrc+size]) {
nDst += size
return nDst, nSrc, transform.ErrShortDst
}
nDst += size
t.prevSpace = false
t.notStart = true
}
nSrc += size
}
return nDst, nSrc, nil
}

153
vendor/golang.org/x/text/secure/precis/options.go generated vendored
View file

@ -1,153 +0,0 @@
// 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.
package precis
import (
"golang.org/x/text/cases"
"golang.org/x/text/language"
"golang.org/x/text/runes"
"golang.org/x/text/transform"
"golang.org/x/text/unicode/norm"
)
// An Option is used to define the behavior and rules of a Profile.
type Option func(*options)
type options struct {
// Preparation options
foldWidth bool
// Enforcement options
asciiLower bool
cases transform.SpanningTransformer
disallow runes.Set
norm transform.SpanningTransformer
additional []func() transform.SpanningTransformer
width transform.SpanningTransformer
disallowEmpty bool
bidiRule bool
// Comparison options
ignorecase bool
}
func getOpts(o ...Option) (res options) {
for _, f := range o {
f(&res)
}
// Using a SpanningTransformer, instead of norm.Form prevents an allocation
// down the road.
if res.norm == nil {
res.norm = norm.NFC
}
return
}
var (
// The IgnoreCase option causes the profile to perform a case insensitive
// comparison during the PRECIS comparison step.
IgnoreCase Option = ignoreCase
// The FoldWidth option causes the profile to map non-canonical wide and
// narrow variants to their decomposition mapping. This is useful for
// profiles that are based on the identifier class which would otherwise
// disallow such characters.
FoldWidth Option = foldWidth
// The DisallowEmpty option causes the enforcement step to return an error if
// the resulting string would be empty.
DisallowEmpty Option = disallowEmpty
// The BidiRule option causes the Bidi Rule defined in RFC 5893 to be
// applied.
BidiRule Option = bidiRule
)
var (
ignoreCase = func(o *options) {
o.ignorecase = true
}
foldWidth = func(o *options) {
o.foldWidth = true
}
disallowEmpty = func(o *options) {
o.disallowEmpty = true
}
bidiRule = func(o *options) {
o.bidiRule = true
}
)
// TODO: move this logic to package transform
type spanWrap struct{ transform.Transformer }
func (s spanWrap) Span(src []byte, atEOF bool) (n int, err error) {
return 0, transform.ErrEndOfSpan
}
// TODO: allow different types? For instance:
// func() transform.Transformer
// func() transform.SpanningTransformer
// func([]byte) bool // validation only
//
// Also, would be great if we could detect if a transformer is reentrant.
// The AdditionalMapping option defines the additional mapping rule for the
// Profile by applying Transformer's in sequence.
func AdditionalMapping(t ...func() transform.Transformer) Option {
return func(o *options) {
for _, f := range t {
sf := func() transform.SpanningTransformer {
return f().(transform.SpanningTransformer)
}
if _, ok := f().(transform.SpanningTransformer); !ok {
sf = func() transform.SpanningTransformer {
return spanWrap{f()}
}
}
o.additional = append(o.additional, sf)
}
}
}
// The Norm option defines a Profile's normalization rule. Defaults to NFC.
func Norm(f norm.Form) Option {
return func(o *options) {
o.norm = f
}
}
// The FoldCase option defines a Profile's case mapping rule. Options can be
// provided to determine the type of case folding used.
func FoldCase(opts ...cases.Option) Option {
return func(o *options) {
o.asciiLower = true
o.cases = cases.Fold(opts...)
}
}
// The LowerCase option defines a Profile's case mapping rule. Options can be
// provided to determine the type of case folding used.
func LowerCase(opts ...cases.Option) Option {
return func(o *options) {
o.asciiLower = true
if len(opts) == 0 {
o.cases = cases.Lower(language.Und, cases.HandleFinalSigma(false))
return
}
opts = append([]cases.Option{cases.HandleFinalSigma(false)}, opts...)
o.cases = cases.Lower(language.Und, opts...)
}
}
// The Disallow option further restricts a Profile's allowed characters beyond
// what is disallowed by the underlying string class.
func Disallow(set runes.Set) Option {
return func(o *options) {
o.disallow = set
}
}

388
vendor/golang.org/x/text/secure/precis/profile.go generated vendored
View file

@ -1,388 +0,0 @@
// 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.
package precis
import (
"bytes"
"errors"
"unicode/utf8"
"golang.org/x/text/cases"
"golang.org/x/text/language"
"golang.org/x/text/runes"
"golang.org/x/text/secure/bidirule"
"golang.org/x/text/transform"
"golang.org/x/text/width"
)
var (
errDisallowedRune = errors.New("precis: disallowed rune encountered")
)
var dpTrie = newDerivedPropertiesTrie(0)
// A Profile represents a set of rules for normalizing and validating strings in
// the PRECIS framework.
type Profile struct {
options
class *class
}
// NewIdentifier creates a new PRECIS profile based on the Identifier string
// class. Profiles created from this class are suitable for use where safety is
// prioritized over expressiveness like network identifiers, user accounts, chat
// rooms, and file names.
func NewIdentifier(opts ...Option) *Profile {
return &Profile{
options: getOpts(opts...),
class: identifier,
}
}
// NewFreeform creates a new PRECIS profile based on the Freeform string class.
// Profiles created from this class are suitable for use where expressiveness is
// prioritized over safety like passwords, and display-elements such as
// nicknames in a chat room.
func NewFreeform(opts ...Option) *Profile {
return &Profile{
options: getOpts(opts...),
class: freeform,
}
}
// NewTransformer creates a new transform.Transformer that performs the PRECIS
// preparation and enforcement steps on the given UTF-8 encoded bytes.
func (p *Profile) NewTransformer() *Transformer {
var ts []transform.Transformer
// These transforms are applied in the order defined in
// https://tools.ietf.org/html/rfc7564#section-7
if p.options.foldWidth {
ts = append(ts, width.Fold)
}
for _, f := range p.options.additional {
ts = append(ts, f())
}
if p.options.cases != nil {
ts = append(ts, p.options.cases)
}
ts = append(ts, p.options.norm)
if p.options.bidiRule {
ts = append(ts, bidirule.New())
}
ts = append(ts, &checker{p: p, allowed: p.Allowed()})
// TODO: Add the disallow empty rule with a dummy transformer?
return &Transformer{transform.Chain(ts...)}
}
var errEmptyString = errors.New("precis: transformation resulted in empty string")
type buffers struct {
src []byte
buf [2][]byte
next int
}
func (b *buffers) apply(t transform.SpanningTransformer) (err error) {
n, err := t.Span(b.src, true)
if err != transform.ErrEndOfSpan {
return err
}
x := b.next & 1
if b.buf[x] == nil {
b.buf[x] = make([]byte, 0, 8+len(b.src)+len(b.src)>>2)
}
span := append(b.buf[x][:0], b.src[:n]...)
b.src, _, err = transform.Append(t, span, b.src[n:])
b.buf[x] = b.src
b.next++
return err
}
// Pre-allocate transformers when possible. In some cases this avoids allocation.
var (
foldWidthT transform.SpanningTransformer = width.Fold
lowerCaseT transform.SpanningTransformer = cases.Lower(language.Und, cases.HandleFinalSigma(false))
)
// TODO: make this a method on profile.
func (b *buffers) enforce(p *Profile, src []byte, comparing bool) (str []byte, err error) {
b.src = src
ascii := true
for _, c := range src {
if c >= utf8.RuneSelf {
ascii = false
break
}
}
// ASCII fast path.
if ascii {
for _, f := range p.options.additional {
if err = b.apply(f()); err != nil {
return nil, err
}
}
switch {
case p.options.asciiLower || (comparing && p.options.ignorecase):
for i, c := range b.src {
if 'A' <= c && c <= 'Z' {
b.src[i] = c ^ 1<<5
}
}
case p.options.cases != nil:
b.apply(p.options.cases)
}
c := checker{p: p}
if _, err := c.span(b.src, true); err != nil {
return nil, err
}
if p.disallow != nil {
for _, c := range b.src {
if p.disallow.Contains(rune(c)) {
return nil, errDisallowedRune
}
}
}
if p.options.disallowEmpty && len(b.src) == 0 {
return nil, errEmptyString
}
return b.src, nil
}
// These transforms are applied in the order defined in
// https://tools.ietf.org/html/rfc7564#section-7
// TODO: allow different width transforms options.
if p.options.foldWidth || (p.options.ignorecase && comparing) {
b.apply(foldWidthT)
}
for _, f := range p.options.additional {
if err = b.apply(f()); err != nil {
return nil, err
}
}
if p.options.cases != nil {
b.apply(p.options.cases)
}
if comparing && p.options.ignorecase {
b.apply(lowerCaseT)
}
b.apply(p.norm)
if p.options.bidiRule && !bidirule.Valid(b.src) {
return nil, bidirule.ErrInvalid
}
c := checker{p: p}
if _, err := c.span(b.src, true); err != nil {
return nil, err
}
if p.disallow != nil {
for i := 0; i < len(b.src); {
r, size := utf8.DecodeRune(b.src[i:])
if p.disallow.Contains(r) {
return nil, errDisallowedRune
}
i += size
}
}
if p.options.disallowEmpty && len(b.src) == 0 {
return nil, errEmptyString
}
return b.src, nil
}
// Append appends the result of applying p to src writing the result to dst.
// It returns an error if the input string is invalid.
func (p *Profile) Append(dst, src []byte) ([]byte, error) {
var buf buffers
b, err := buf.enforce(p, src, false)
if err != nil {
return nil, err
}
return append(dst, b...), nil
}
func processBytes(p *Profile, b []byte, key bool) ([]byte, error) {
var buf buffers
b, err := buf.enforce(p, b, key)
if err != nil {
return nil, err
}
if buf.next == 0 {
c := make([]byte, len(b))
copy(c, b)
return c, nil
}
return b, nil
}
// Bytes returns a new byte slice with the result of applying the profile to b.
func (p *Profile) Bytes(b []byte) ([]byte, error) {
return processBytes(p, b, false)
}
// AppendCompareKey appends the result of applying p to src (including any
// optional rules to make strings comparable or useful in a map key such as
// applying lowercasing) writing the result to dst. It returns an error if the
// input string is invalid.
func (p *Profile) AppendCompareKey(dst, src []byte) ([]byte, error) {
var buf buffers
b, err := buf.enforce(p, src, true)
if err != nil {
return nil, err
}
return append(dst, b...), nil
}
func processString(p *Profile, s string, key bool) (string, error) {
var buf buffers
b, err := buf.enforce(p, []byte(s), key)
if err != nil {
return "", err
}
return string(b), nil
}
// String returns a string with the result of applying the profile to s.
func (p *Profile) String(s string) (string, error) {
return processString(p, s, false)
}
// CompareKey returns a string that can be used for comparison, hashing, or
// collation.
func (p *Profile) CompareKey(s string) (string, error) {
return processString(p, s, true)
}
// Compare enforces both strings, and then compares them for bit-string identity
// (byte-for-byte equality). If either string cannot be enforced, the comparison
// is false.
func (p *Profile) Compare(a, b string) bool {
var buf buffers
akey, err := buf.enforce(p, []byte(a), true)
if err != nil {
return false
}
buf = buffers{}
bkey, err := buf.enforce(p, []byte(b), true)
if err != nil {
return false
}
return bytes.Compare(akey, bkey) == 0
}
// Allowed returns a runes.Set containing every rune that is a member of the
// underlying profile's string class and not disallowed by any profile specific
// rules.
func (p *Profile) Allowed() runes.Set {
if p.options.disallow != nil {
return runes.Predicate(func(r rune) bool {
return p.class.Contains(r) && !p.options.disallow.Contains(r)
})
}
return p.class
}
type checker struct {
p *Profile
allowed runes.Set
beforeBits catBitmap
termBits catBitmap
acceptBits catBitmap
}
func (c *checker) Reset() {
c.beforeBits = 0
c.termBits = 0
c.acceptBits = 0
}
func (c *checker) span(src []byte, atEOF bool) (n int, err error) {
for n < len(src) {
e, sz := dpTrie.lookup(src[n:])
d := categoryTransitions[category(e&catMask)]
if sz == 0 {
if !atEOF {
return n, transform.ErrShortSrc
}
return n, errDisallowedRune
}
if property(e) < c.p.class.validFrom {
if d.rule == nil {
return n, errDisallowedRune
}
doLookAhead, err := d.rule(c.beforeBits)
if err != nil {
return n, err
}
if doLookAhead {
c.beforeBits &= d.keep
c.beforeBits |= d.set
// We may still have a lookahead rule which we will require to
// complete (by checking termBits == 0) before setting the new
// bits.
if c.termBits != 0 && (!c.checkLookahead() || c.termBits == 0) {
return n, err
}
c.termBits = d.term
c.acceptBits = d.accept
n += sz
continue
}
}
c.beforeBits &= d.keep
c.beforeBits |= d.set
if c.termBits != 0 && !c.checkLookahead() {
return n, errContext
}
n += sz
}
if m := c.beforeBits >> finalShift; c.beforeBits&m != m || c.termBits != 0 {
err = errContext
}
return n, err
}
func (c *checker) checkLookahead() bool {
switch {
case c.beforeBits&c.termBits != 0:
c.termBits = 0
c.acceptBits = 0
case c.beforeBits&c.acceptBits != 0:
default:
return false
}
return true
}
// TODO: we may get rid of this transform if transform.Chain understands
// something like a Spanner interface.
func (c checker) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
short := false
if len(dst) < len(src) {
src = src[:len(dst)]
atEOF = false
short = true
}
nSrc, err = c.span(src, atEOF)
nDst = copy(dst, src[:nSrc])
if short && (err == transform.ErrShortSrc || err == nil) {
err = transform.ErrShortDst
}
return nDst, nSrc, err
}

78
vendor/golang.org/x/text/secure/precis/profiles.go generated vendored
View file

@ -1,78 +0,0 @@
// 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.
package precis
import (
"unicode"
"golang.org/x/text/runes"
"golang.org/x/text/transform"
"golang.org/x/text/unicode/norm"
)
var (
// Implements the Nickname profile specified in RFC 7700.
// The nickname profile is not idempotent and may need to be applied multiple
// times before being used for comparisons.
Nickname *Profile = nickname
// Implements the UsernameCaseMapped profile specified in RFC 7613.
UsernameCaseMapped *Profile = usernameCaseMap
// Implements the UsernameCasePreserved profile specified in RFC 7613.
UsernameCasePreserved *Profile = usernameNoCaseMap
// Implements the OpaqueString profile defined in RFC 7613 for passwords and other secure labels.
OpaqueString *Profile = opaquestring
)
var (
nickname = &Profile{
options: getOpts(
AdditionalMapping(func() transform.Transformer {
return &nickAdditionalMapping{}
}),
IgnoreCase,
Norm(norm.NFKC),
DisallowEmpty,
),
class: freeform,
}
usernameCaseMap = &Profile{
options: getOpts(
FoldWidth,
LowerCase(),
Norm(norm.NFC),
BidiRule,
),
class: identifier,
}
usernameNoCaseMap = &Profile{
options: getOpts(
FoldWidth,
Norm(norm.NFC),
BidiRule,
),
class: identifier,
}
opaquestring = &Profile{
options: getOpts(
AdditionalMapping(func() transform.Transformer {
return mapSpaces
}),
Norm(norm.NFC),
DisallowEmpty,
),
class: freeform,
}
)
// mapSpaces is a shared value of a runes.Map transformer.
var mapSpaces transform.Transformer = runes.Map(func(r rune) rune {
if unicode.Is(unicode.Zs, r) {
return ' '
}
return r
})

3788
vendor/golang.org/x/text/secure/precis/tables.go generated vendored
View file

File diff suppressed because it is too large Load diff

32
vendor/golang.org/x/text/secure/precis/transformer.go generated vendored
View file

@ -1,32 +0,0 @@
// 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.
package precis
import "golang.org/x/text/transform"
// Transformer implements the transform.Transformer interface.
type Transformer struct {
t transform.Transformer
}
// Reset implements the transform.Transformer interface.
func (t Transformer) Reset() { t.t.Reset() }
// Transform implements the transform.Transformer interface.
func (t Transformer) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
return t.t.Transform(dst, src, atEOF)
}
// Bytes returns a new byte slice with the result of applying t to b.
func (t Transformer) Bytes(b []byte) []byte {
b, _, _ = transform.Bytes(t, b)
return b
}
// String returns a string with the result of applying t to s.
func (t Transformer) String(s string) string {
s, _, _ = transform.String(t, s)
return s
}

64
vendor/golang.org/x/text/secure/precis/trieval.go generated vendored
View file

@ -1,64 +0,0 @@
// Code generated by running "go generate" in golang.org/x/text. DO NOT EDIT.
package precis
// entry is the entry of a trie table
// 7..6 property (unassigned, disallowed, maybe, valid)
// 5..0 category
type entry uint8
const (
propShift = 6
propMask = 0xc0
catMask = 0x3f
)
func (e entry) property() property { return property(e & propMask) }
func (e entry) category() category { return category(e & catMask) }
type property uint8
// The order of these constants matter. A Profile may consider runes to be
// allowed either from pValid or idDisOrFreePVal.
const (
unassigned property = iota << propShift
disallowed
idDisOrFreePVal // disallowed for Identifier, pValid for FreeForm
pValid
)
// compute permutations of all properties and specialCategories.
type category uint8
const (
other category = iota
// Special rune types
joiningL
joiningD
joiningT
joiningR
viramaModifier
viramaJoinT // Virama + JoiningT
latinSmallL // U+006c
greek
greekJoinT // Greek + JoiningT
hebrew
hebrewJoinT // Hebrew + JoiningT
japanese // hirigana, katakana, han
// Special rune types associated with contextual rules defined in
// https://tools.ietf.org/html/rfc5892#appendix-A.
// ContextO
zeroWidthNonJoiner // rule 1
zeroWidthJoiner // rule 2
// ContextJ
middleDot // rule 3
greekLowerNumeralSign // rule 4
hebrewPreceding // rule 5 and 6
katakanaMiddleDot // rule 7
arabicIndicDigit // rule 8
extendedArabicIndicDigit // rule 9
numCategories
)

115
vendor/golang.org/x/text/width/gen.go generated vendored
View file

@ -1,115 +0,0 @@
// 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
// This program generates the trie for width operations. The generated table
// includes width category information as well as the normalization mappings.
package main
import (
"bytes"
"fmt"
"io"
"log"
"math"
"unicode/utf8"
"golang.org/x/text/internal/gen"
"golang.org/x/text/internal/triegen"
)
// See gen_common.go for flags.
func main() {
gen.Init()
genTables()
genTests()
gen.Repackage("gen_trieval.go", "trieval.go", "width")
gen.Repackage("gen_common.go", "common_test.go", "width")
}
func genTables() {
t := triegen.NewTrie("width")
// fold and inverse mappings. See mapComment for a description of the format
// of each entry. Add dummy value to make an index of 0 mean no mapping.
inverse := [][4]byte{{}}
mapping := map[[4]byte]int{[4]byte{}: 0}
getWidthData(func(r rune, tag elem, alt rune) {
idx := 0
if alt != 0 {
var buf [4]byte
buf[0] = byte(utf8.EncodeRune(buf[1:], alt))
s := string(r)
buf[buf[0]] ^= s[len(s)-1]
var ok bool
if idx, ok = mapping[buf]; !ok {
idx = len(mapping)
if idx > math.MaxUint8 {
log.Fatalf("Index %d does not fit in a byte.", idx)
}
mapping[buf] = idx
inverse = append(inverse, buf)
}
}
t.Insert(r, uint64(tag|elem(idx)))
})
w := &bytes.Buffer{}
gen.WriteUnicodeVersion(w)
sz, err := t.Gen(w)
if err != nil {
log.Fatal(err)
}
sz += writeMappings(w, inverse)
fmt.Fprintf(w, "// Total table size %d bytes (%dKiB)\n", sz, sz/1024)
gen.WriteGoFile(*outputFile, "width", w.Bytes())
}
const inverseDataComment = `
// inverseData contains 4-byte entries of the following format:
// <length> <modified UTF-8-encoded rune> <0 padding>
// The last byte of the UTF-8-encoded rune is xor-ed with the last byte of the
// UTF-8 encoding of the original rune. Mappings often have the following
// pattern:
// Ａ -> A (U+FF21 -> U+0041)
// Ｂ -> B (U+FF22 -> U+0042)
// ...
// By xor-ing the last byte the same entry can be shared by many mappings. This
// reduces the total number of distinct entries by about two thirds.
// The resulting entry for the aforementioned mappings is
// { 0x01, 0xE0, 0x00, 0x00 }
// Using this entry to map U+FF21 (UTF-8 [EF BC A1]), we get
// E0 ^ A1 = 41.
// Similarly, for U+FF22 (UTF-8 [EF BC A2]), we get
// E0 ^ A2 = 42.
// Note that because of the xor-ing, the byte sequence stored in the entry is
// not valid UTF-8.`
func writeMappings(w io.Writer, data [][4]byte) int {
fmt.Fprintln(w, inverseDataComment)
fmt.Fprintf(w, "var inverseData = [%d][4]byte{\n", len(data))
for _, x := range data {
fmt.Fprintf(w, "{ 0x%02x, 0x%02x, 0x%02x, 0x%02x },\n", x[0], x[1], x[2], x[3])
}
fmt.Fprintln(w, "}")
return len(data) * 4
}
func genTests() {
w := &bytes.Buffer{}
fmt.Fprintf(w, "\nvar mapRunes = map[rune]struct{r rune; e elem}{\n")
getWidthData(func(r rune, tag elem, alt rune) {
if alt != 0 {
fmt.Fprintf(w, "\t0x%X: {0x%X, 0x%X},\n", r, alt, tag)
}
})
fmt.Fprintln(w, "}")
gen.WriteGoFile("runes_test.go", "width", w.Bytes())
}

96
vendor/golang.org/x/text/width/gen_common.go generated vendored
View file

@ -1,96 +0,0 @@
// 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 code is shared between the main code generator and the test code.
import (
"flag"
"log"
"strconv"
"strings"
"golang.org/x/text/internal/gen"
"golang.org/x/text/internal/ucd"
)
var (
outputFile = flag.String("out", "tables.go", "output file")
)
var typeMap = map[string]elem{
"A": tagAmbiguous,
"N": tagNeutral,
"Na": tagNarrow,
"W": tagWide,
"F": tagFullwidth,
"H": tagHalfwidth,
}
// getWidthData calls f for every entry for which it is defined.
//
// f may be called multiple times for the same rune. The last call to f is the
// correct value. f is not called for all runes. The default tag type is
// Neutral.
func getWidthData(f func(r rune, tag elem, alt rune)) {
// Set the default values for Unified Ideographs. In line with Annex 11,
// we encode full ranges instead of the defined runes in Unified_Ideograph.
for _, b := range []struct{ lo, hi rune }{
{0x4E00, 0x9FFF}, // the CJK Unified Ideographs block,
{0x3400, 0x4DBF}, // the CJK Unified Ideographs Externsion A block,
{0xF900, 0xFAFF}, // the CJK Compatibility Ideographs block,
{0x20000, 0x2FFFF}, // the Supplementary Ideographic Plane,
{0x30000, 0x3FFFF}, // the Tertiary Ideographic Plane,
} {
for r := b.lo; r <= b.hi; r++ {
f(r, tagWide, 0)
}
}
inverse := map[rune]rune{}
maps := map[string]bool{
"<wide>": true,
"<narrow>": true,
}
// We cannot reuse package norm's decomposition, as we need an unexpanded
// decomposition. We make use of the opportunity to verify that the
// decomposition type is as expected.
ucd.Parse(gen.OpenUCDFile("UnicodeData.txt"), func(p *ucd.Parser) {
r := p.Rune(0)
s := strings.SplitN(p.String(ucd.DecompMapping), " ", 2)
if !maps[s[0]] {
return
}
x, err := strconv.ParseUint(s[1], 16, 32)
if err != nil {
log.Fatalf("Error parsing rune %q", s[1])
}
if inverse[r] != 0 || inverse[rune(x)] != 0 {
log.Fatalf("Circular dependency in mapping between %U and %U", r, x)
}
inverse[r] = rune(x)
inverse[rune(x)] = r
})
// <rune range>;<type>
ucd.Parse(gen.OpenUCDFile("EastAsianWidth.txt"), func(p *ucd.Parser) {
tag, ok := typeMap[p.String(1)]
if !ok {
log.Fatalf("Unknown width type %q", p.String(1))
}
r := p.Rune(0)
alt, ok := inverse[r]
if tag == tagFullwidth || tag == tagHalfwidth && r != wonSign {
tag |= tagNeedsFold
if !ok {
log.Fatalf("Narrow or wide rune %U has no decomposition", r)
}
}
f(r, tag, alt)
})
}

34
vendor/golang.org/x/text/width/gen_trieval.go generated vendored
View file

@ -1,34 +0,0 @@
// 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
// elem is an entry of the width trie. The high byte is used to encode the type
// of the rune. The low byte is used to store the index to a mapping entry in
// the inverseData array.
type elem uint16
const (
tagNeutral elem = iota << typeShift
tagAmbiguous
tagWide
tagNarrow
tagFullwidth
tagHalfwidth
)
const (
numTypeBits = 3
typeShift = 16 - numTypeBits
// tagNeedsFold is true for all fullwidth and halfwidth runes except for
// the Won sign U+20A9.
tagNeedsFold = 0x1000
// The Korean Won sign is halfwidth, but SHOULD NOT be mapped to a wide
// variant.
wonSign rune = 0x20A9
)

16
vendor/golang.org/x/text/width/kind_string.go generated vendored
View file

@ -1,16 +0,0 @@
// Code generated by "stringer -type=Kind"; DO NOT EDIT
package width
import "fmt"
const _Kind_name = "NeutralEastAsianAmbiguousEastAsianWideEastAsianNarrowEastAsianFullwidthEastAsianHalfwidth"
var _Kind_index = [...]uint8{0, 7, 25, 38, 53, 71, 89}
func (i Kind) String() string {
if i < 0 || i >= Kind(len(_Kind_index)-1) {
return fmt.Sprintf("Kind(%d)", i)
}
return _Kind_name[_Kind_index[i]:_Kind_index[i+1]]
}

1284
vendor/golang.org/x/text/width/tables.go generated vendored
View file

File diff suppressed because it is too large Load diff

239
vendor/golang.org/x/text/width/transform.go generated vendored
View file

@ -1,239 +0,0 @@
// 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.
package width
import (
"unicode/utf8"
"golang.org/x/text/transform"
)
type foldTransform struct {
transform.NopResetter
}
func (foldTransform) Span(src []byte, atEOF bool) (n int, err error) {
for n < len(src) {
if src[n] < utf8.RuneSelf {
// ASCII fast path.
for n++; n < len(src) && src[n] < utf8.RuneSelf; n++ {
}
continue
}
v, size := trie.lookup(src[n:])
if size == 0 { // incomplete UTF-8 encoding
if !atEOF {
err = transform.ErrShortSrc
} else {
n = len(src)
}
break
}
if elem(v)&tagNeedsFold != 0 {
err = transform.ErrEndOfSpan
break
}
n += size
}
return n, err
}
func (foldTransform) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
for nSrc < len(src) {
if src[nSrc] < utf8.RuneSelf {
// ASCII fast path.
start, end := nSrc, len(src)
if d := len(dst) - nDst; d < end-start {
end = nSrc + d
}
for nSrc++; nSrc < end && src[nSrc] < utf8.RuneSelf; nSrc++ {
}
n := copy(dst[nDst:], src[start:nSrc])
if nDst += n; nDst == len(dst) {
nSrc = start + n
if nSrc == len(src) {
return nDst, nSrc, nil
}
if src[nSrc] < utf8.RuneSelf {
return nDst, nSrc, transform.ErrShortDst
}
}
continue
}
v, size := trie.lookup(src[nSrc:])
if size == 0 { // incomplete UTF-8 encoding
if !atEOF {
return nDst, nSrc, transform.ErrShortSrc
}
size = 1 // gobble 1 byte
}
if elem(v)&tagNeedsFold == 0 {
if size != copy(dst[nDst:], src[nSrc:nSrc+size]) {
return nDst, nSrc, transform.ErrShortDst
}
nDst += size
} else {
data := inverseData[byte(v)]
if len(dst)-nDst < int(data[0]) {
return nDst, nSrc, transform.ErrShortDst
}
i := 1
for end := int(data[0]); i < end; i++ {
dst[nDst] = data[i]
nDst++
}
dst[nDst] = data[i] ^ src[nSrc+size-1]
nDst++
}
nSrc += size
}
return nDst, nSrc, nil
}
type narrowTransform struct {
transform.NopResetter
}
func (narrowTransform) Span(src []byte, atEOF bool) (n int, err error) {
for n < len(src) {
if src[n] < utf8.RuneSelf {
// ASCII fast path.
for n++; n < len(src) && src[n] < utf8.RuneSelf; n++ {
}
continue
}
v, size := trie.lookup(src[n:])
if size == 0 { // incomplete UTF-8 encoding
if !atEOF {
err = transform.ErrShortSrc
} else {
n = len(src)
}
break
}
if k := elem(v).kind(); byte(v) == 0 || k != EastAsianFullwidth && k != EastAsianWide && k != EastAsianAmbiguous {
} else {
err = transform.ErrEndOfSpan
break
}
n += size
}
return n, err
}
func (narrowTransform) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
for nSrc < len(src) {
if src[nSrc] < utf8.RuneSelf {
// ASCII fast path.
start, end := nSrc, len(src)
if d := len(dst) - nDst; d < end-start {
end = nSrc + d
}
for nSrc++; nSrc < end && src[nSrc] < utf8.RuneSelf; nSrc++ {
}
n := copy(dst[nDst:], src[start:nSrc])
if nDst += n; nDst == len(dst) {
nSrc = start + n
if nSrc == len(src) {
return nDst, nSrc, nil
}
if src[nSrc] < utf8.RuneSelf {
return nDst, nSrc, transform.ErrShortDst
}
}
continue
}
v, size := trie.lookup(src[nSrc:])
if size == 0 { // incomplete UTF-8 encoding
if !atEOF {
return nDst, nSrc, transform.ErrShortSrc
}
size = 1 // gobble 1 byte
}
if k := elem(v).kind(); byte(v) == 0 || k != EastAsianFullwidth && k != EastAsianWide && k != EastAsianAmbiguous {
if size != copy(dst[nDst:], src[nSrc:nSrc+size]) {
return nDst, nSrc, transform.ErrShortDst
}
nDst += size
} else {
data := inverseData[byte(v)]
if len(dst)-nDst < int(data[0]) {
return nDst, nSrc, transform.ErrShortDst
}
i := 1
for end := int(data[0]); i < end; i++ {
dst[nDst] = data[i]
nDst++
}
dst[nDst] = data[i] ^ src[nSrc+size-1]
nDst++
}
nSrc += size
}
return nDst, nSrc, nil
}
type wideTransform struct {
transform.NopResetter
}
func (wideTransform) Span(src []byte, atEOF bool) (n int, err error) {
for n < len(src) {
// TODO: Consider ASCII fast path. Special-casing ASCII handling can
// reduce the ns/op of BenchmarkWideASCII by about 30%. This is probably
// not enough to warrant the extra code and complexity.
v, size := trie.lookup(src[n:])
if size == 0 { // incomplete UTF-8 encoding
if !atEOF {
err = transform.ErrShortSrc
} else {
n = len(src)
}
break
}
if k := elem(v).kind(); byte(v) == 0 || k != EastAsianHalfwidth && k != EastAsianNarrow {
} else {
err = transform.ErrEndOfSpan
break
}
n += size
}
return n, err
}
func (wideTransform) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
for nSrc < len(src) {
// TODO: Consider ASCII fast path. Special-casing ASCII handling can
// reduce the ns/op of BenchmarkWideASCII by about 30%. This is probably
// not enough to warrant the extra code and complexity.
v, size := trie.lookup(src[nSrc:])
if size == 0 { // incomplete UTF-8 encoding
if !atEOF {
return nDst, nSrc, transform.ErrShortSrc
}
size = 1 // gobble 1 byte
}
if k := elem(v).kind(); byte(v) == 0 || k != EastAsianHalfwidth && k != EastAsianNarrow {
if size != copy(dst[nDst:], src[nSrc:nSrc+size]) {
return nDst, nSrc, transform.ErrShortDst
}
nDst += size
} else {
data := inverseData[byte(v)]
if len(dst)-nDst < int(data[0]) {
return nDst, nSrc, transform.ErrShortDst
}
i := 1
for end := int(data[0]); i < end; i++ {
dst[nDst] = data[i]
nDst++
}
dst[nDst] = data[i] ^ src[nSrc+size-1]
nDst++
}
nSrc += size
}
return nDst, nSrc, nil
}

30
vendor/golang.org/x/text/width/trieval.go generated vendored
View file

@ -1,30 +0,0 @@
// Code generated by running "go generate" in golang.org/x/text. DO NOT EDIT.
package width
// elem is an entry of the width trie. The high byte is used to encode the type
// of the rune. The low byte is used to store the index to a mapping entry in
// the inverseData array.
type elem uint16
const (
tagNeutral elem = iota << typeShift
tagAmbiguous
tagWide
tagNarrow
tagFullwidth
tagHalfwidth
)
const (
numTypeBits = 3
typeShift = 16 - numTypeBits
// tagNeedsFold is true for all fullwidth and halfwidth runes except for
// the Won sign U+20A9.
tagNeedsFold = 0x1000
// The Korean Won sign is halfwidth, but SHOULD NOT be mapped to a wide
// variant.
wonSign rune = 0x20A9
)

206
vendor/golang.org/x/text/width/width.go generated vendored
View file

@ -1,206 +0,0 @@
// 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.
//go:generate stringer -type=Kind
//go:generate go run gen.go gen_common.go gen_trieval.go
// Package width provides functionality for handling different widths in text.
//
// Wide characters behave like ideographs; they tend to allow line breaks after
// each character and remain upright in vertical text layout. Narrow characters
// are kept together in words or runs that are rotated sideways in vertical text
// layout.
//
// For more information, see http://unicode.org/reports/tr11/.
package width // import "golang.org/x/text/width"
import (
"unicode/utf8"
"golang.org/x/text/transform"
)
// TODO
// 1) Reduce table size by compressing blocks.
// 2) API proposition for computing display length
// (approximation, fixed pitch only).
// 3) Implement display length.
// Kind indicates the type of width property as defined in http://unicode.org/reports/tr11/.
type Kind int
const (
// Neutral characters do not occur in legacy East Asian character sets.
Neutral Kind = iota
// EastAsianAmbiguous characters that can be sometimes wide and sometimes
// narrow and require additional information not contained in the character
// code to further resolve their width.
EastAsianAmbiguous
// EastAsianWide characters are wide in its usual form. They occur only in
// the context of East Asian typography. These runes may have explicit
// halfwidth counterparts.
EastAsianWide
// EastAsianNarrow characters are narrow in its usual form. They often have
// fullwidth counterparts.
EastAsianNarrow
// Note: there exist Narrow runes that do not have fullwidth or wide
// counterparts, despite what the definition says (e.g. U+27E6).
// EastAsianFullwidth characters have a compatibility decompositions of type
// wide that map to a narrow counterpart.
EastAsianFullwidth
// EastAsianHalfwidth characters have a compatibility decomposition of type
// narrow that map to a wide or ambiguous counterpart, plus U+20A9 ₩ WON
// SIGN.
EastAsianHalfwidth
// Note: there exist runes that have a halfwidth counterparts but that are
// classified as Ambiguous, rather than wide (e.g. U+2190).
)
// TODO: the generated tries need to return size 1 for invalid runes for the
// width to be computed correctly (each byte should render width 1)
var trie = newWidthTrie(0)
// Lookup reports the Properties of the first rune in b and the number of bytes
// of its UTF-8 encoding.
func Lookup(b []byte) (p Properties, size int) {
v, sz := trie.lookup(b)
return Properties{elem(v), b[sz-1]}, sz
}
// LookupString reports the Properties of the first rune in s and the number of
// bytes of its UTF-8 encoding.
func LookupString(s string) (p Properties, size int) {
v, sz := trie.lookupString(s)
return Properties{elem(v), s[sz-1]}, sz
}
// LookupRune reports the Properties of rune r.
func LookupRune(r rune) Properties {
var buf [4]byte
n := utf8.EncodeRune(buf[:], r)
v, _ := trie.lookup(buf[:n])
last := byte(r)
if r >= utf8.RuneSelf {
last = 0x80 + byte(r&0x3f)
}
return Properties{elem(v), last}
}
// Properties provides access to width properties of a rune.
type Properties struct {
elem elem
last byte
}
func (e elem) kind() Kind {
return Kind(e >> typeShift)
}
// Kind returns the Kind of a rune as defined in Unicode TR #11.
// See http://unicode.org/reports/tr11/ for more details.
func (p Properties) Kind() Kind {
return p.elem.kind()
}
// Folded returns the folded variant of a rune or 0 if the rune is canonical.
func (p Properties) Folded() rune {
if p.elem&tagNeedsFold != 0 {
buf := inverseData[byte(p.elem)]
buf[buf[0]] ^= p.last
r, _ := utf8.DecodeRune(buf[1 : 1+buf[0]])
return r
}
return 0
}
// Narrow returns the narrow variant of a rune or 0 if the rune is already
// narrow or doesn't have a narrow variant.
func (p Properties) Narrow() rune {
if k := p.elem.kind(); byte(p.elem) != 0 && (k == EastAsianFullwidth || k == EastAsianWide || k == EastAsianAmbiguous) {
buf := inverseData[byte(p.elem)]
buf[buf[0]] ^= p.last
r, _ := utf8.DecodeRune(buf[1 : 1+buf[0]])
return r
}
return 0
}
// Wide returns the wide variant of a rune or 0 if the rune is already
// wide or doesn't have a wide variant.
func (p Properties) Wide() rune {
if k := p.elem.kind(); byte(p.elem) != 0 && (k == EastAsianHalfwidth || k == EastAsianNarrow) {
buf := inverseData[byte(p.elem)]
buf[buf[0]] ^= p.last
r, _ := utf8.DecodeRune(buf[1 : 1+buf[0]])
return r
}
return 0
}
// TODO for Properties:
// - Add Fullwidth/Halfwidth or Inverted methods for computing variants
// mapping.
// - Add width information (including information on non-spacing runes).
// Transformer implements the transform.Transformer interface.
type Transformer struct {
t transform.SpanningTransformer
}
// Reset implements the transform.Transformer interface.
func (t Transformer) Reset() { t.t.Reset() }
// Transform implements the transform.Transformer interface.
func (t Transformer) Transform(dst, src []byte, atEOF bool) (nDst, nSrc int, err error) {
return t.t.Transform(dst, src, atEOF)
}
// Span implements the transform.SpanningTransformer interface.
func (t Transformer) Span(src []byte, atEOF bool) (n int, err error) {
return t.t.Span(src, atEOF)
}
// Bytes returns a new byte slice with the result of applying t to b.
func (t Transformer) Bytes(b []byte) []byte {
b, _, _ = transform.Bytes(t, b)
return b
}
// String returns a string with the result of applying t to s.
func (t Transformer) String(s string) string {
s, _, _ = transform.String(t, s)
return s
}
var (
// Fold is a transform that maps all runes to their canonical width.
//
// Note that the NFKC and NFKD transforms in golang.org/x/text/unicode/norm
// provide a more generic folding mechanism.
Fold Transformer = Transformer{foldTransform{}}
// Widen is a transform that maps runes to their wide variant, if
// available.
Widen Transformer = Transformer{wideTransform{}}
// Narrow is a transform that maps runes to their narrow variant, if
// available.
Narrow Transformer = Transformer{narrowTransform{}}
)
// TODO: Consider the following options:
// - Treat Ambiguous runes that have a halfwidth counterpart as wide, or some
// generalized variant of this.
// - Consider a wide Won character to be the default width (or some generalized
// variant of this).
// - Filter the set of characters that gets converted (the preferred approach is
// to allow applying filters to transforms).