/// <summary>
/// Reset Builder's internal state with the given language tag
/// </summary>
public InternalLocaleBuilder SetLanguageTag(LanguageTag langtag)
{
Clear();
if (langtag.Extlangs.Count > 0)
{
_language = langtag.Extlangs[0];
}
else
{
string language = langtag.Language;
if (!language.Equals(LanguageTag.Undetermined))
{
_language = language;
}
}
_script = langtag.Script;
_region = langtag.Region;
IList <string> bcpVariants = langtag.Variants;
if (bcpVariants.Count > 0)
{
StringBuilder var = new StringBuilder(bcpVariants[0]);
for (int i = 1; i < bcpVariants.Count; i++)
{
var.Append(BaseLocale.Separator).Append(bcpVariants[i]);
}
_variant = var.ToString();
}
SetExtensions(langtag.Extensions, langtag.PrivateUse);
return(this);
}
/// <summary>
/// Reset Builder's internal state with the given language tag
/// </summary>
public InternalLocaleBuilder SetLanguageTag(LanguageTag langtag)
{
Clear();
if (langtag.GetExtlangs().Count > 0)
{
_language = langtag.GetExtlangs()[0];
}
else
{
string language = langtag.GetLanguage();
if (!language.Equals(LanguageTag.UNDETERMINED))
{
_language = language;
}
}
_script = langtag.GetScript();
_region = langtag.GetRegion();
IList <string> bcpVariants = langtag.GetVariants();
if (bcpVariants.Count > 0)
{
StringBuilder var = new StringBuilder(bcpVariants[0]);
for (int i = 1; i < bcpVariants.Count; i++)
{
var.Append(BaseLocale.SEP).Append(bcpVariants[i]);
}
_variant = var.ToString();
}
SetExtensions(langtag.GetExtensions(), langtag.GetPrivateuse());
return(this);
}
private static string ToID(IDictionary <char, Extension> map)
{
StringBuilder buf = new StringBuilder();
Extension privuse = null;
foreach (var entry in map)
{
char singleton = entry.Key;
Extension extension = entry.Value;
if (LanguageTag.IsPrivateusePrefixChar(singleton))
{
privuse = extension;
}
else
{
if (buf.Length > 0)
{
buf.Append(LanguageTag.Separator);
}
buf.Append(extension);
}
}
if (privuse != null)
{
if (buf.Length > 0)
{
buf.Append(LanguageTag.Separator);
}
buf.Append(privuse);
}
return(buf.ToString());
}
/// <summary>
/// Check if the given variant subtags separated by the given
/// separator(s) are valid.
/// </summary>
private int CheckVariants(string variants, string sep)
{
StringTokenEnumerator itr = new StringTokenEnumerator(variants, sep);
while (itr.MoveNext())
{
string s = itr.Current;
if (!LanguageTag.IsVariant(s))
{
return(itr.CurrentStart);
}
}
return(-1);
}
public InternalLocaleBuilder SetRegion(string region)
{
if (region == null || region.Length == 0)
{
_region = "";
}
else
{
if (!LanguageTag.IsRegion(region))
{
throw new FormatException("Ill-formed region: " + region /*, 0*/);
}
_region = region;
}
return(this);
}
public InternalLocaleBuilder SetScript(string script)
{
if (script == null || script.Length == 0)
{
_script = "";
}
else
{
if (!LanguageTag.IsScript(script))
{
throw new FormatException("Ill-formed script: " + script /*, 0*/);
}
_script = script;
}
return(this);
}
public InternalLocaleBuilder SetLanguage(string language)
{
if (language == null || language.Length == 0)
{
_language = "";
}
else
{
if (!LanguageTag.IsLanguage(language))
{
throw new FormatException("Ill-formed language: " + language /*, 0*/);
}
_language = language;
}
return(this);
}
public InternalLocaleBuilder SetLocale(BaseLocale @base, LocaleExtensions extensions)
{
string language = @base.GetLanguage();
string script = @base.GetScript();
string region = @base.GetRegion();
string variant = @base.GetVariant();
// ICU4N TODO: Remove ?
if (JDKIMPL)
{
// Special backward compatibility support
// Exception 1 - ja_JP_JP
if (language.Equals("ja") && region.Equals("JP") && variant.Equals("JP"))
{
// When locale ja_JP_JP is created, ca-japanese is always there.
// The builder ignores the variant "JP"
Debug.Assert("japanese".Equals(extensions.GetUnicodeLocaleType("ca")));
variant = "";
}
// Exception 2 - th_TH_TH
else if (language.Equals("th") && region.Equals("TH") && variant.Equals("TH"))
{
// When locale th_TH_TH is created, nu-thai is always there.
// The builder ignores the variant "TH"
Debug.Assert("thai".Equals(extensions.GetUnicodeLocaleType("nu")));
variant = "";
}
// Exception 3 - no_NO_NY
else if (language.Equals("no") && region.Equals("NO") && variant.Equals("NY")) // ICU4N TODO: Fix this handling for .NET (no-NO is not reliable across platforms)
{
// no_NO_NY is a valid locale and used by Java 6 or older versions.
// The build ignores the variant "NY" and change the language to "nn".
language = "nn";
variant = "";
}
}
// Validate base locale fields before updating internal state.
// LocaleExtensions always store validated/canonicalized values,
// so no checks are necessary.
if (language.Length > 0 && !LanguageTag.IsLanguage(language))
{
throw new FormatException("Ill-formed language: " + language);
}
if (script.Length > 0 && !LanguageTag.IsScript(script))
{
throw new FormatException("Ill-formed script: " + script);
}
if (region.Length > 0 && !LanguageTag.IsRegion(region))
{
throw new FormatException("Ill-formed region: " + region); // ICU4N TODO: Port LocaleSyntaxException (instead of FormatException)
}
if (variant.Length > 0)
{
int errIdx = CheckVariants(variant, BaseLocale.Separator);
if (errIdx != -1)
{
throw new FormatException("Ill-formed variant: " + variant /*, errIdx*/);
}
}
// The input locale is validated at this point.
// Now, updating builder's internal fields.
_language = language;
_script = script;
_region = region;
_variant = variant;
ClearExtensions();
var extKeys = (extensions == null) ? null : extensions.Keys;
if (extKeys != null)
{
// map extensions back to builder's internal format
foreach (char key in extKeys)
{
Extension e = extensions.GetExtension(key);
if (e is UnicodeLocaleExtension ue)
{
foreach (string uatr in ue.UnicodeLocaleAttributes)
{
if (_uattributes == null)
{
_uattributes = new HashSet <CaseInsensitiveString>(/*4*/);
}
_uattributes.Add(new CaseInsensitiveString(uatr));
}
foreach (string ukey in ue.UnicodeLocaleKeys)
{
if (_ukeywords == null)
{
_ukeywords = new Dictionary <CaseInsensitiveString, string>(4);
}
_ukeywords[new CaseInsensitiveString(ukey)] = ue.GetUnicodeLocaleType(ukey);
}
}
else
{
if (_extensions == null)
{
_extensions = new Dictionary <CaseInsensitiveChar, string>(4);
}
_extensions[new CaseInsensitiveChar(key)] = e.Value;
}
}
}
return(this);
}
/// <summary>
/// Set extension/private subtags in a single string representation
/// </summary>
public InternalLocaleBuilder SetExtensions(string subtags)
{
if (subtags == null || subtags.Length == 0)
{
ClearExtensions();
return(this);
}
subtags = subtags.Replace(BaseLocale.Separator, LanguageTag.Separator);
StringTokenEnumerator itr = new StringTokenEnumerator(subtags, LanguageTag.Separator);
List <string> extensions = null;
string privateuse = null;
int parsed = 0;
int start;
// Move to first element
itr.MoveNext();
// Make a list of extension subtags
while (!itr.IsDone)
{
string s = itr.Current;
if (LanguageTag.IsExtensionSingleton(s))
{
start = itr.CurrentStart;
string singleton = s;
StringBuilder sb = new StringBuilder(singleton);
itr.MoveNext();
while (!itr.IsDone)
{
s = itr.Current;
if (LanguageTag.IsExtensionSubtag(s))
{
sb.Append(LanguageTag.Separator).Append(s);
parsed = itr.CurrentEnd;
}
else
{
break;
}
itr.MoveNext();
}
if (parsed < start)
{
throw new FormatException("Incomplete extension '" + singleton + "'" /*, start*/);
}
if (extensions == null)
{
extensions = new List <string>(4);
}
extensions.Add(sb.ToString());
}
else
{
break;
}
}
if (!itr.IsDone)
{
string s = itr.Current;
if (LanguageTag.IsPrivateusePrefix(s))
{
start = itr.CurrentStart;
StringBuilder sb = new StringBuilder(s);
itr.MoveNext();
while (!itr.IsDone)
{
s = itr.Current;
if (!LanguageTag.IsPrivateuseSubtag(s))
{
break;
}
sb.Append(LanguageTag.Separator).Append(s);
parsed = itr.CurrentEnd;
itr.MoveNext();
}
if (parsed <= start)
{
throw new FormatException("Incomplete privateuse:" + subtags.Substring(start) /*, start*/);
}
else
{
privateuse = sb.ToString();
}
}
}
if (!itr.IsDone)
{
throw new FormatException("Ill-formed extension subtags:" + subtags.Substring(itr.CurrentStart) /*, itr.CurrentStart*/);
}
return(SetExtensions(extensions, privateuse));
}
public InternalLocaleBuilder SetExtension(char singleton, string value)
{
// validate key
bool isBcpPrivateuse = LanguageTag.IsPrivateusePrefixChar(singleton);
if (!isBcpPrivateuse && !LanguageTag.IsExtensionSingletonChar(singleton))
{
throw new FormatException("Ill-formed extension key: " + singleton);
}
bool remove = (value == null || value.Length == 0);
CaseInsensitiveChar key = new CaseInsensitiveChar(singleton);
if (remove)
{
if (UnicodeLocaleExtension.IsSingletonChar(key.Value))
{
// clear entire Unicode locale extension
if (_uattributes != null)
{
_uattributes.Clear();
}
if (_ukeywords != null)
{
_ukeywords.Clear();
}
}
else
{
if (_extensions != null && _extensions.ContainsKey(key))
{
_extensions.Remove(key);
}
}
}
else
{
// validate value
string val = value.Replace(BaseLocale.Separator, LanguageTag.Separator);
StringTokenEnumerator itr = new StringTokenEnumerator(val, LanguageTag.Separator);
while (itr.MoveNext())
{
string s = itr.Current;
bool validSubtag;
if (isBcpPrivateuse)
{
validSubtag = LanguageTag.IsPrivateuseSubtag(s);
}
else
{
validSubtag = LanguageTag.IsExtensionSubtag(s);
}
if (!validSubtag)
{
throw new FormatException("Ill-formed extension value: " + s /*, itr.CurrentStart*/);
}
}
if (UnicodeLocaleExtension.IsSingletonChar(key.Value))
{
SetUnicodeLocaleExtension(val);
}
else
{
if (_extensions == null)
{
_extensions = new Dictionary <CaseInsensitiveChar, string>(4);
}
_extensions[key] = val;
}
}
return(this);
}
public static LanguageTag ParseLocale(BaseLocale baseLocale, LocaleExtensions localeExtensions)
{
LanguageTag tag = new LanguageTag();
string language = baseLocale.GetLanguage();
string script = baseLocale.GetScript();
string region = baseLocale.GetRegion();
string variant = baseLocale.GetVariant();
bool hasSubtag = false;
string privuseVar = null; // store ill-formed variant subtags
if (language.Length > 0 && IsLanguage(language))
{
// Convert a deprecated language code used by Java to
// a new code
if (language.Equals("iw"))
{
language = "he";
}
else if (language.Equals("ji"))
{
language = "yi";
}
else if (language.Equals("in"))
{
language = "id";
}
tag._language = language;
}
if (script.Length > 0 && IsScript(script))
{
tag._script = CanonicalizeScript(script);
hasSubtag = true;
}
if (region.Length > 0 && IsRegion(region))
{
tag._region = CanonicalizeRegion(region);
hasSubtag = true;
}
// ICU4N TODO: Remove ?
if (JDKIMPL)
{
// Special handling for no_NO_NY - use nn_NO for language tag
if (tag._language.Equals("no") && tag._region.Equals("NO") && variant.Equals("NY")) // ICU4N TODO: Fix this handling for .NET (no-NO is not reliable across platforms)
{
tag._language = "nn";
variant = "";
}
}
if (variant.Length > 0)
{
List <string> variants = null;
StringTokenEnumerator varitr = new StringTokenEnumerator(variant, BaseLocale.SEP);
while (varitr.MoveNext())
{
string var = varitr.Current;
if (!IsVariant(var))
{
break;
}
if (variants == null)
{
variants = new List <string>();
}
if (JDKIMPL)
{
variants.Add(var); // Do not canonicalize!
}
else
{
variants.Add(CanonicalizeVariant(var));
}
}
if (variants != null)
{
tag._variants = variants;
hasSubtag = true;
}
if (!varitr.IsDone)
{
// ill-formed variant subtags
StringBuilder buf = new StringBuilder();
while (!varitr.IsDone)
{
string prvv = varitr.Current;
if (!IsPrivateuseSubtag(prvv))
{
// cannot use private use subtag - truncated
break;
}
if (buf.Length > 0)
{
buf.Append(SEP);
}
if (!JDKIMPL)
{
prvv = AsciiUtil.ToLowerString(prvv);
}
buf.Append(prvv);
varitr.MoveNext();
}
if (buf.Length > 0)
{
privuseVar = buf.ToString();
}
}
}
List <string> extensions = null;
string privateuse = null;
var locextKeys = localeExtensions.Keys;
foreach (char locextKey in locextKeys)
{
Extension ext = localeExtensions.GetExtension(locextKey);
if (IsPrivateusePrefixChar(locextKey))
{
privateuse = ext.Value;
}
else
{
if (extensions == null)
{
extensions = new List <string>();
}
extensions.Add(locextKey.ToString() + SEP + ext.Value);
}
}
if (extensions != null)
{
tag._extensions = extensions;
hasSubtag = true;
}
// append ill-formed variant subtags to private use
if (privuseVar != null)
{
if (privateuse == null)
{
privateuse = PRIVUSE_VARIANT_PREFIX + SEP + privuseVar;
}
else
{
privateuse = privateuse + SEP + PRIVUSE_VARIANT_PREFIX + SEP + privuseVar.Replace(BaseLocale.SEP, SEP);
}
}
if (privateuse != null)
{
tag._privateuse = privateuse;
}
if (tag._language.Length == 0 && (hasSubtag || privateuse == null))
{
// use lang "und" when 1) no language is available AND
// 2) any of other subtags other than private use are available or
// no private use tag is available
tag._language = UNDETERMINED;
}
return(tag);
}
/// <summary>
/// BNF in RFC5464
/// </summary>
/// <remarks>
/// Language-Tag = langtag ; normal language tags
/// / privateuse ; private use tag
/// / grandfathered ; grandfathered tags
///
///
/// langtag = language
/// ["-" script]
/// ["-" region]
/// *("-" variant)
/// *("-" extension)
/// ["-" privateuse]
///
/// language = 2*3ALPHA ; shortest ISO 639 code
/// ["-" extlang] ; sometimes followed by
/// ; extended language subtags
/// / 4ALPHA ; or reserved for future use
/// / 5*8ALPHA ; or registered language subtag
///
/// extlang = 3ALPHA ; selected ISO 639 codes
/// *2("-" 3ALPHA) ; permanently reserved
///
/// script = 4ALPHA ; ISO 15924 code
///
/// region = 2ALPHA ; ISO 3166-1 code
/// / 3DIGIT ; UN M.49 code
///
/// variant = 5*8alphanum ; registered variants
/// / (DIGIT 3alphanum)
///
/// extension = singleton 1*("-" (2*8alphanum))
///
/// ; Single alphanumerics
/// ; "x" reserved for private use
/// singleton = DIGIT ; 0 - 9
/// / %x41-57 ; A - W
/// / %x59-5A ; Y - Z
/// / %x61-77 ; a - w
/// / %x79-7A ; y - z
///
/// privateuse = "x" 1*("-" (1*8alphanum))
/// </remarks>
public static LanguageTag Parse(string languageTag, ParseStatus sts)
{
if (sts == null)
{
sts = new ParseStatus();
}
else
{
sts.Reset();
}
StringTokenEnumerator itr;
bool isGrandfathered = false;
// Check if the tag is grandfathered
string[] gfmap;
if (GRANDFATHERED.TryGetValue(new AsciiUtil.CaseInsensitiveKey(languageTag), out gfmap) && gfmap != null)
{
// use preferred mapping
itr = new StringTokenEnumerator(gfmap[1], SEP);
isGrandfathered = true;
}
else
{
itr = new StringTokenEnumerator(languageTag, SEP);
}
// ICU4N: Move to the first element
itr.MoveNext();
LanguageTag tag = new LanguageTag();
// langtag must start with either language or privateuse
if (tag.ParseLanguage(itr, sts))
{
tag.ParseExtlangs(itr, sts);
tag.ParseScript(itr, sts);
tag.ParseRegion(itr, sts);
tag.ParseVariants(itr, sts);
tag.ParseExtensions(itr, sts);
}
tag.ParsePrivateuse(itr, sts);
if (isGrandfathered)
{
// Grandfathered tag is replaced with a well-formed tag above.
// However, the parsed length must be the original tag length.
Debug.Assert(itr.IsDone);
Debug.Assert(!sts.IsError);
sts.ParseLength = languageTag.Length;
}
else if (!itr.IsDone && !sts.IsError)
{
string s = itr.Current;
sts.ErrorIndex = itr.CurrentStart;
if (s.Length == 0)
{
sts.ErrorMessage = "Empty subtag";
}
else
{
sts.ErrorMessage = "Invalid subtag: " + s;
}
}
return(tag);
}
/// <summary>
/// Internal constructor, only used by <see cref="InternalLocaleBuilder"/>.
/// </summary>
internal LocaleExtensions(IDictionary <CaseInsensitiveChar, string> extensions,
ISet <CaseInsensitiveString> uattributes, IDictionary <CaseInsensitiveString, string> ukeywords)
{
bool hasExtension = (extensions != null && extensions.Count > 0);
bool hasUAttributes = (uattributes != null && uattributes.Count > 0);
bool hasUKeywords = (ukeywords != null && ukeywords.Count > 0);
if (!hasExtension && !hasUAttributes && !hasUKeywords)
{
_map = EmptyMap;
_id = "";
return;
}
// Build extension map
_map = new JCG.SortedDictionary <char, Extension>();
if (hasExtension)
{
foreach (var ext in extensions)
{
char key = AsciiUtil.ToLower(ext.Key.Value);
string value = ext.Value;
if (LanguageTag.IsPrivateusePrefixChar(key))
{
// we need to exclude special variant in privuateuse, e.g. "x-abc-lvariant-DEF"
value = InternalLocaleBuilder.RemovePrivateuseVariant(value);
if (value == null)
{
continue;
}
}
Extension e = new Extension(key, AsciiUtil.ToLower(value));
_map[key] = e;
}
}
if (hasUAttributes || hasUKeywords)
{
JCG.SortedSet <string> uaset = null;
JCG.SortedDictionary <string, string> ukmap = null;
if (hasUAttributes)
{
uaset = new JCG.SortedSet <string>(StringComparer.Ordinal);
foreach (CaseInsensitiveString cis in uattributes)
{
uaset.Add(AsciiUtil.ToLower(cis.Value));
}
}
if (hasUKeywords)
{
ukmap = new JCG.SortedDictionary <string, string>(StringComparer.Ordinal);
foreach (var kwd in ukeywords)
{
string key = AsciiUtil.ToLower(kwd.Key.Value);
string type = AsciiUtil.ToLower(kwd.Value);
ukmap[key] = type;
}
}
UnicodeLocaleExtension ule = new UnicodeLocaleExtension(uaset, ukmap);
_map[UnicodeLocaleExtension.Singleton] = ule;
}
if (_map.Count == 0)
{
// this could happen when only privuateuse with special variant
_map = EmptyMap;
_id = "";
}
else
{
_id = ToID(_map);
}
}
public static bool IsValidKey(char c)
{
return(LanguageTag.IsExtensionSingletonChar(c) || LanguageTag.IsPrivateusePrefixChar(c));
}
