/// <summary> /// This is used to move one part if appropriate from the 'from' subtag to private use. /// Alternatively, if any part is appropriate for the tag according to the test predicate, /// it is moved to the first position (unless it follows an x- and keepStandardPartInPrivateUse is true). /// If we didn't find a valid part, but did move something, insert standardPrivateCode at the START of "from". /// As a side effect, this method may remove non-alphanumeric characters from the from tag. /// (I don't like having such a side effect, but it seems necessary to produce the desired behavior). /// </summary> private void MoveFirstPartToPrivateUseIfNecessary(SubTag from, Func <string, bool> test, string standardPrivatePart, bool keepStandardPartInPrivateUse) { string part = from.AllParts.FirstOrDefault(); if (part == null) { return; // nothing to move. } if (test(part)) { return; // no need to move, it is a valid code for its slot. } if (MoveStandardPartToStart(from, test, keepStandardPartInPrivateUse)) { return; } // If we exit this loop we need to move the first part to private use. // But first strip illegal characters since that may leave nothing to move, // or at least nothing of the first part we would otherwise move. // We do NOT want to do this BEFORE looking for good parts, because (for example) if we have a // region code like U!S-gb, we want to detect 'gb' as a good region code and keep that, // rather than fixing U!S to US and then choosing to keep that. from.RemoveNonAlphaNumericCharacters(); // But, now we should scan again. If cleaning out bad characters resulted in a good code, // let's put it in the main part of the tag rather than private-use. if (MoveStandardPartToStart(from, test, keepStandardPartInPrivateUse)) { return; } // OK, no good parts left. We will move the first part that is not an X. part = FirstNonXPart(from.AllParts); if (part == null) { return; } _privateUseSubTag.AddToSubtag(part); from.RemoveParts(part); from.InsertAtStartOfSubtag(standardPrivatePart); }
/// <summary> /// Cleans the tag. /// </summary> public void Clean() { // Migrate legacy ISO3 language codes to IANA 2 letter language codes, if there's a match. // Do this before we look for valid codes, otherwise the 3-letter ones come up as invalid and // get moved to private use. However, only do this to languages not identified as private-use. if (!Language.StartsWith("x-", StringComparison.OrdinalIgnoreCase)) { string migrateFrom = ""; string migrateTo = ""; foreach (string part in _languageSubTag.AllParts) { if (part.Equals("x", StringComparison.OrdinalIgnoreCase)) { break; // don't migrate language code parts already explicitly marked private-use. } if (string.IsNullOrEmpty(migrateFrom)) { LanguageSubtag language; if (StandardSubtags.TryGetLanguageFromIso3Code(part, out language) && language.Code != language.Iso3Code) { migrateFrom = part; migrateTo = language.Code; } } } if (!String.IsNullOrEmpty(migrateFrom)) { _languageSubTag.RemoveParts(migrateFrom); _languageSubTag.AddToSubtag(migrateTo); } } // The very next thing, before anything else gets moved to private use, is to move the parts whose position we // care about to the appropriate position in the private use section. // In the process we may remove anything non-alphanumeric, since otherwise we may move a marker that later // disappears (pathologically). MoveFirstPartToPrivateUseIfNecessary(_languageSubTag, StandardSubtags.IsValidIso639LanguageCode, "qaa", true); MoveFirstPartToPrivateUseIfNecessary(_scriptSubTag, StandardSubtags.IsValidIso15924ScriptCode, "Qaaa", false); MoveFirstPartToPrivateUseIfNecessary(_regionSubTag, StandardSubtags.IsValidIso3166RegionCode, "QM", false); //This fixes a bug where the LdmlAdaptorV1 was writing out Zxxx as part of the variant to mark an audio writing system if (_variantSubTag.Contains(WellKnownSubtags.AudioScript)) { MoveTagsMatching(_variantSubTag, _scriptSubTag, tag => tag.Equals(WellKnownSubtags.AudioScript)); _privateUseSubTag.AddToSubtag(WellKnownSubtags.AudioPrivateUse); } // Fixes various legacy problems. if (Language.Equals("cmn", StringComparison.OrdinalIgnoreCase)) { Language = "zh"; } if (Language.Equals("pes", StringComparison.OrdinalIgnoreCase)) { Language = "fa"; } if (Language.Equals("arb", StringComparison.OrdinalIgnoreCase)) { Language = "ar"; } if (Language.Equals("zh", StringComparison.OrdinalIgnoreCase) && String.IsNullOrEmpty(Region)) { Region = "CN"; } // If the language tag contains an x- , then move the string behind the x- to private use MovePartsToPrivateUseIfNecessary(_languageSubTag); // Move script, region, and variant present in the langauge tag to their proper subtag. MoveTagsMatching(_languageSubTag, _scriptSubTag, StandardSubtags.IsValidIso15924ScriptCode, StandardSubtags.IsValidIso639LanguageCode); MoveTagsMatching(_languageSubTag, _regionSubTag, StandardSubtags.IsValidIso3166RegionCode, StandardSubtags.IsValidIso639LanguageCode); MoveTagsMatching(_languageSubTag, _variantSubTag, StandardSubtags.IsValidRegisteredVariantCode, StandardSubtags.IsValidIso639LanguageCode); // Move all other tags that don't belong to the private use subtag. //keep track of everything that we moved var tempSubTag = new SubTag(); MoveTagsMatching(_languageSubTag, tempSubTag, tag => !StandardSubtags.IsValidIso639LanguageCode(tag)); //place all the moved parts in private use. foreach (var part in tempSubTag.AllParts) { _privateUseSubTag.AddToSubtag(part); //if it looks like we moved a custom script set the subtag to mark that we've moved it if (_scriptSubTag.IsEmpty && part.Length == 4 && //potential custom script tag !WellKnownSubtags.IpaPhonemicPrivateUse.EndsWith(part) && !WellKnownSubtags.IpaPhoneticPrivateUse.EndsWith(part)) { _scriptSubTag = new SubTag("Qaaa"); } } MoveTagsMatching(_scriptSubTag, _privateUseSubTag, tag => !StandardSubtags.IsValidIso15924ScriptCode(tag)); MoveTagsMatching(_regionSubTag, _privateUseSubTag, tag => !StandardSubtags.IsValidIso3166RegionCode(tag)); MoveTagsMatching(_variantSubTag, _privateUseSubTag, tag => !StandardSubtags.IsValidRegisteredVariantCode(tag)); _languageSubTag.KeepFirstAndMoveRemainderTo(_privateUseSubTag); _scriptSubTag.KeepFirstAndMoveRemainderTo(_privateUseSubTag); _regionSubTag.KeepFirstAndMoveRemainderTo(_privateUseSubTag); if (_privateUseSubTag.Contains(WellKnownSubtags.AudioPrivateUse)) { // Move every tag that's not a Zxxx to private use if (!_scriptSubTag.IsEmpty && !_scriptSubTag.Contains(WellKnownSubtags.AudioScript)) { MoveTagsMatching(_scriptSubTag, _privateUseSubTag, tag => !_privateUseSubTag.Contains(tag)); } // If we don't have a Zxxx already, set it. This protects tags already present, but with unusual case if (!_scriptSubTag.Contains(WellKnownSubtags.AudioScript)) { _scriptSubTag = new SubTag(WellKnownSubtags.AudioScript); } } //These two methods may produce duplicates that will subsequently be removed. Do we care? - TA 29/3/2011 _privateUseSubTag.RemoveNonAlphaNumericCharacters(); _privateUseSubTag.TruncatePartsToNumCharacters(8); _variantSubTag.RemoveDuplicates(); _privateUseSubTag.RemoveDuplicates(); // Any 'x' in the other tags will have arrived in the privateUse tag, so remove them. _privateUseSubTag.RemoveParts("x"); // if language is empty, we need to add qaa, unless only a privateUse is present (e.g. x-blah is a valid rfc5646 tag) if ((_languageSubTag.IsEmpty && (!_scriptSubTag.IsEmpty || !_regionSubTag.IsEmpty || !_variantSubTag.IsEmpty)) || (_languageSubTag.IsEmpty && _scriptSubTag.IsEmpty && _regionSubTag.IsEmpty && _variantSubTag.IsEmpty && _privateUseSubTag.IsEmpty)) { _languageSubTag.AddToSubtag("qaa"); } // Two more legacy problems. We don't allow -etic or -emic without fonipa, so insert if needed. // If it has some other standard variant we won't be able to fix it...not sure what the right answer would be. // At least we catch the more common case. foreach (string part in _privateUseSubTag.AllParts) { if (string.IsNullOrEmpty(Variant) && (part.Equals("etic", StringComparison.OrdinalIgnoreCase) || part.Equals("emic", StringComparison.OrdinalIgnoreCase))) { Variant = "fonipa"; } } }
/// <summary> /// This is used to move one part if appropriate from the 'from' subtag to private use. /// Alternatively, if any part is appropriate for the tag according to the test predicate, /// it is moved to the first position (unless it follows an x- and keepStandardPartInPrivateUse is true). /// If we didn't find a valid part, but did move something, insert standardPrivateCode at the START of "from". /// As a side effect, this method may remove non-alphanumeric characters from the from tag. /// (I don't like having such a side effect, but it seems necessary to produce the desired behavior). /// </summary> private void MoveFirstPartToPrivateUseIfNecessary(SubTag from, Func<string, bool> test, string standardPrivatePart, bool keepStandardPartInPrivateUse) { var part = from.AllParts.FirstOrDefault(); if (part == null) return; // nothing to move. if (test(part)) return; // no need to move, it is a valid code for its slot. if (MoveStandardPartToStart(from, test, keepStandardPartInPrivateUse)) return; // If we exit this loop we need to move the first part to private use. // But first strip illegal characters since that may leave nothing to move, // or at least nothing of the first part we would otherwise move. // We do NOT want to do this BEFORE looking for good parts, because (for example) if we have a // region code like U!S-gb, we want to detect 'gb' as a good region code and keep that, // rather than fixing U!S to US and then choosing to keep that. from.RemoveNonAlphaNumericCharacters(); // But, now we should scan again. If cleaning out bad characters resulted in a good code, // let's put it in the main part of the tag rather than private-use. if (MoveStandardPartToStart(from, test, keepStandardPartInPrivateUse)) return; // OK, no good parts left. We will move the first part that is not an X. part = FirstNonXPart(from.AllParts); if (part == null) return; _privateUseSubTag.AddToSubtag(part); from.RemoveParts(part); from.InsertAtStartOfSubtag(standardPrivatePart); }