Version history
Where to find the released versions of each Data Standardizer package.
There is no changelog or release-notes file in this repository. For the version history of released packages, see each package’s listing on NuGet:
Data Standardizer packages on NuGet
Each package page on NuGet lists its published versions and release dates.
Removed in v2.0
Version 2.0 removes the legacy DataStandardizer.BCP47, DataStandardizer.File, DataStandardizer.ISO15924, DataStandardizer.ISO3166, DataStandardizer.ISO4217, DataStandardizer.ISO639 and DataStandardizer.UNM49 packages. See Removed in v2.0 for the package that replaces each one and the types that were renamed. It also raises the lowest target of DataStandardizer.Communication, DataStandardizer.LanguageTag and DataStandardizer.Money from netstandard1.0 to netstandard1.1. It changes the ExtendedLanguageSubtags, VariantSubtags and ExtensionSubtags properties of Bcp47LanguageTag, and the Subtags property of Bcp47KeyedSubtag, from arrays to IReadOnlyList<T>, and removes the DataStandardizer.LanguageTag members that took a match timeout, and the builder’s UsingRegionSubtag overloads that took a UN M49 country code enum.
Bcp47LanguageTag.PrimaryLanguageSubtag is now a string?. On a private use tag such as x-private, or a grandfathered tag such as i-klingon, it is null instead of the empty string, so code that tests for an empty string should test for null instead.
Creating a Bcp47LanguageTag without a subtag registry now checks that the tag is well-formed and nothing else, so it accepts tags such as zz and de-1901-1901 that version 1.x rejected. IsWellFormedLanguageTagString is replaced by IsWellFormed and IsValid, and the ISO code lists are an explicit opt-in, which now rejects three-letter codes such as eng. Check*Subtag and the builder’s string steps check only the shape of a subtag. See what creating a language tag checks.
In DataStandardizer.File.CSV, ICsvFileLine is no longer an IOrderedDictionary. It is an ordered collection of fields with string names, enumerated as KeyValuePair<string, object?> items, with typed Keys and Values lists. See the File.CSV line model.
A File.CSV mapper registered with RegisterMapper now applies only to the reader or writer it is registered on, not to every reader and writer of the record type, and the protected members of CsvFileIoBase<TRecordLine> and CsvFileCacheRepositoryBase are removed. CsvMapperCache.Clear() clears the cached mappers built from attributes. See File.CSV mapper registration.
A File.CSV reader now parses a mapped numeric, date, time, bool, Guid or enum property that has no converter, instead of leaving it at its default value. A field that the built-in parse rejects, or that a TypeConverter throws for, goes to the bad-value handler or raises a CsvFileException. An exception from a CsvFieldConvertFromString delegate still propagates unchanged. The FieldIndex a validator and the bad-value handler get is now the field’s position in the line. See File.CSV conversion of mapped properties.
When a bad-value handler is set, a File.CSV reader now reads the last field of a record that breaks the quoting rules as an empty value, as it already did for every other field, instead of keeping its raw text. See File.CSV invalid last field.
A File.CSV record line property that calls GetPropertyValue<T> now throws an InvalidCastException when its field holds a value of another type, such as the text a reader stores for a field that no mapping converts, instead of returning the default value of its type. See File.CSV typed property values.
SubtagRegistry moves to the DataStandardizer.LanguageTag namespace, and the types of its records move to DataStandardizer.LanguageTag.Registry. See where the registry types live.
SubtagRegistry.CreateFrom… now requires CRLF at the end of every line of the registry by default, as RFC 5646 §3.1.1 requires. IANA publishes the registry with a line feed alone at the end of each line, so load it with SubtagRegistryLoadOptions.AcceptBareLineFeed, which the CreateFrom… methods take. SubtagRegistry.ToString() and the records’ ToString() now write CRLF at the end of every line, the last one included, instead of Environment.NewLine with no line break after the last line.
In DataStandardizer.LanguageTag, Subtag Registry records are now immutable, with their fields parsed once when the registry is loaded:
ISubtagRegistryRecordis no longer anIList<Tuple<string, object>>. Read a record’s raw fields from itsFieldsproperty, a read-only list ofKeyValuePair<string, string>(KeyandValuein place ofItem1andItem2).SubtagRegistryTagRecordBase.Typeis aSubtagRegistryRecordTypeinstead of a string: compare withSubtagRegistryRecordType.Languagerather than"language".DescriptionandPrefixareIReadOnlyList<string>instead of arrays, andCommentsis anIReadOnlyList<string>with everyCommentsfield instead of a string with only the first.- A date that isn’t in the
yyyy-MM-ddform reads as the defaultDateTime(ornullforDeprecated) instead of throwingFormatException. SubtagRegistryFieldAttribute, the protectedGetPropertyValue<T>()andGetPropertyValues<T>()helpers, and the publicSubtagRegistryFileDateRecordconstructor are removed. The record classes are sealed and their constructors internal, so records come only from a loadedSubtagRegistry.
SubtagRegistry gains TryGetSubtag, IsGrandfathered, FileDate and a list of the records of each type (Languages, ExtendedLanguages, Scripts, Regions, Variants, GrandfatheredTags and RedundantTags), so code that filtered the records by type can use those instead.
How the docs track versions
This documentation reflects the current state of the packages rather than any single released version. Where a feature is tied to a particular target framework or release, the relevant page carries a short “Applies to” note instead of the docs being split into versioned sets. The Platform support page is a good example of this convention in action.