2016-11-03 18:16:01 -04:00
|
|
|
// 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"
|
|
|
|
"strconv"
|
|
|
|
"strings"
|
|
|
|
|
2019-06-18 22:14:15 -04:00
|
|
|
"golang.org/x/text/internal/language"
|
2016-11-03 18:16:01 -04:00
|
|
|
)
|
|
|
|
|
|
|
|
// 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.
|
2019-06-18 22:14:15 -04:00
|
|
|
type ValueError interface {
|
|
|
|
error
|
2016-11-03 18:16:01 -04:00
|
|
|
|
2019-06-18 22:14:15 -04:00
|
|
|
// Subtag returns the subtag for which the error occurred.
|
|
|
|
Subtag() string
|
2016-11-03 18:16:01 -04:00
|
|
|
}
|
|
|
|
|
|
|
|
// 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
|
2019-06-18 22:14:15 -04:00
|
|
|
// https://www.unicode.org/reports/tr35/#Unicode_Language_and_Locale_Identifiers.
|
2016-11-03 18:16:01 -04:00
|
|
|
// 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
|
2019-06-18 22:14:15 -04:00
|
|
|
// https://www.unicode.org/reports/tr35/#Unicode_Language_and_Locale_Identifiers.
|
|
|
|
// The resulting tag is canonicalized using the canonicalization type c.
|
2016-11-03 18:16:01 -04:00
|
|
|
func (c CanonType) Parse(s string) (t Tag, err error) {
|
2019-06-18 22:14:15 -04:00
|
|
|
tt, err := language.Parse(s)
|
|
|
|
if err != nil {
|
|
|
|
return makeTag(tt), err
|
2016-11-03 18:16:01 -04:00
|
|
|
}
|
2019-06-18 22:14:15 -04:00
|
|
|
tt, changed := canonicalize(c, tt)
|
2016-11-03 18:16:01 -04:00
|
|
|
if changed {
|
2019-06-18 22:14:15 -04:00
|
|
|
tt.RemakeString()
|
2016-11-03 18:16:01 -04:00
|
|
|
}
|
2019-06-18 22:14:15 -04:00
|
|
|
return makeTag(tt), err
|
2016-11-03 18:16:01 -04:00
|
|
|
}
|
|
|
|
|
|
|
|
// 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
|
2019-06-18 22:14:15 -04:00
|
|
|
// will replace the former. For -u extensions, though, the key-type pairs are
|
|
|
|
// added, where later values overwrite older ones. 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.
|
2016-11-03 18:16:01 -04:00
|
|
|
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
|
2019-06-18 22:14:15 -04:00
|
|
|
// will replace the former. For -u extensions, though, the key-type pairs are
|
|
|
|
// added, where later values overwrite older ones. 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.
|
2016-11-03 18:16:01 -04:00
|
|
|
func (c CanonType) Compose(part ...interface{}) (t Tag, err error) {
|
2019-06-18 22:14:15 -04:00
|
|
|
var b language.Builder
|
|
|
|
if err = update(&b, part...); err != nil {
|
2016-11-03 18:16:01 -04:00
|
|
|
return und, err
|
|
|
|
}
|
2019-06-18 22:14:15 -04:00
|
|
|
b.Tag, _ = canonicalize(c, b.Tag)
|
|
|
|
return makeTag(b.Make()), err
|
2016-11-03 18:16:01 -04:00
|
|
|
}
|
|
|
|
|
|
|
|
var errInvalidArgument = errors.New("invalid Extension or Variant")
|
|
|
|
|
2019-06-18 22:14:15 -04:00
|
|
|
func update(b *language.Builder, part ...interface{}) (err error) {
|
2016-11-03 18:16:01 -04:00
|
|
|
for _, x := range part {
|
|
|
|
switch v := x.(type) {
|
|
|
|
case Tag:
|
2019-06-18 22:14:15 -04:00
|
|
|
b.SetTag(v.tag())
|
2016-11-03 18:16:01 -04:00
|
|
|
case Base:
|
2019-06-18 22:14:15 -04:00
|
|
|
b.Tag.LangID = v.langID
|
2016-11-03 18:16:01 -04:00
|
|
|
case Script:
|
2019-06-18 22:14:15 -04:00
|
|
|
b.Tag.ScriptID = v.scriptID
|
2016-11-03 18:16:01 -04:00
|
|
|
case Region:
|
2019-06-18 22:14:15 -04:00
|
|
|
b.Tag.RegionID = v.regionID
|
2016-11-03 18:16:01 -04:00
|
|
|
case Variant:
|
2019-06-18 22:14:15 -04:00
|
|
|
if v.variant == "" {
|
|
|
|
err = errInvalidArgument
|
|
|
|
break
|
2016-11-03 18:16:01 -04:00
|
|
|
}
|
2019-06-18 22:14:15 -04:00
|
|
|
b.AddVariant(v.variant)
|
2016-11-03 18:16:01 -04:00
|
|
|
case Extension:
|
2019-06-18 22:14:15 -04:00
|
|
|
if v.s == "" {
|
|
|
|
err = errInvalidArgument
|
|
|
|
break
|
2016-11-03 18:16:01 -04:00
|
|
|
}
|
2019-06-18 22:14:15 -04:00
|
|
|
b.SetExt(v.s)
|
2016-11-03 18:16:01 -04:00
|
|
|
case []Variant:
|
2019-06-18 22:14:15 -04:00
|
|
|
b.ClearVariants()
|
|
|
|
for _, v := range v {
|
|
|
|
b.AddVariant(v.variant)
|
2016-11-03 18:16:01 -04:00
|
|
|
}
|
|
|
|
case []Extension:
|
2019-06-18 22:14:15 -04:00
|
|
|
b.ClearExtensions()
|
2016-11-03 18:16:01 -04:00
|
|
|
for _, e := range v {
|
2019-06-18 22:14:15 -04:00
|
|
|
b.SetExt(e.s)
|
2016-11-03 18:16:01 -04:00
|
|
|
}
|
|
|
|
// TODO: support parsing of raw strings based on morphology or just extensions?
|
|
|
|
case error:
|
2019-06-18 22:14:15 -04:00
|
|
|
if v != nil {
|
|
|
|
err = v
|
2016-11-03 18:16:01 -04:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2019-06-18 22:14:15 -04:00
|
|
|
return
|
2016-11-03 18:16:01 -04:00
|
|
|
}
|
|
|
|
|
|
|
|
var errInvalidWeight = errors.New("ParseAcceptLanguage: invalid weight")
|
|
|
|
|
2019-03-27 07:15:23 -04:00
|
|
|
// ParseAcceptLanguage parses the contents of an Accept-Language header as
|
2016-11-03 18:16:01 -04:00
|
|
|
// 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
|
|
|
|
}
|
2019-06-18 22:14:15 -04:00
|
|
|
t = makeTag(language.Tag{LangID: id})
|
2016-11-03 18:16:01 -04:00
|
|
|
}
|
|
|
|
|
|
|
|
// 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), ""
|
|
|
|
}
|
|
|
|
|
2019-06-18 22:14:15 -04:00
|
|
|
// Add hack mapping to deal with a small number of cases that occur
|
2016-11-03 18:16:01 -04:00
|
|
|
// in Accept-Language (with reasonable frequency).
|
2019-06-18 22:14:15 -04:00
|
|
|
var acceptFallback = map[string]language.Language{
|
2016-11-03 18:16:01 -04:00
|
|
|
"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]
|
|
|
|
}
|