Class HyphenationRegistry

Namespace

NetPdf.Hyphenation

Assembly

NetPdf.Text.dll

The public extension seam for language hyphenation. The core ships American-English patterns (pre-registered as "en"); the optional NetPdf.Languages.* NuGet packages register additional languages here (via an explicit Register(…) call, which the packs also drive from a [ModuleInitializer]) by handing over TeX-style Liang pattern text + an optional exception list. Patterns registered through this seam are reachable via TryHyphenate(string, ReadOnlySpan<char>, out int[], int, int) and IsRegistered(string).

public static class HyphenationRegistry

Inheritance

object

HyphenationRegistry

Inherited Members

object.Equals(object)

object.Equals(object, object)

object.GetHashCode()

object.GetType()

object.MemberwiseClone()

object.ReferenceEquals(object, object)

object.ToString()

Remarks

Language keys. Keys are normalized to the primary subtag, lower-cased (BCP-47 de-DE → de, en-GB → en), so a pack registering "de" serves all German locales. A tag with no primary subtag after normalization (e.g. "-DE" or "-") is rejected by Register(string, string, string?).

Thread-safety. Registration and lookup are concurrent-safe. Registering a language that is already present replaces it (last registration wins) — a pack can override the built-in English.

Layout routing. The block layout pass resolves the hyphens: auto hyphenator from a block's effective HTML lang (nearest lang/xml:lang up the ancestor chain) through this registry, via the internal ResolveOrDefault(string) seam. So a language a pack registered here is used automatically during rendering. An unregistered/unknown language falls back to the bundled English hyphenator. Still a follow-up: per-run routing (a single block that mixes languages uses the block language for all of it) and UAX #29 tokenization.

Methods

IsRegistered(string)

Whether a hyphenator is registered for language (by primary subtag).

public static bool IsRegistered(string language)

Parameters

language string

Returns

bool

Register(string, string, string?)

Register (or replace) the hyphenator for language from TeX-style Liang patternBlock text (whitespace/newline-separated patterns like .aus1, TeX % comments stripped) and an optional exceptionBlock of explicit hyphenations (whitespace-separated words like Sil-ben-tren-nung). Called by the NetPdf.Languages.* packs on load; may also be called directly to supply custom patterns.

public static void Register(string language, string patternBlock, string? exceptionBlock = null)

Parameters

language string

patternBlock string

exceptionBlock string

RegisterNoHyphenation(string)

Register language as one that does NOT soft-hyphenate — a script whose line breaking is handled elsewhere rather than by inserting hyphens (CJK: per-character breaking, UAX #14; Arabic: kashida/tatweel justification). This registers a no-op hyphenator, so hyphens: auto resolves to zero break points for the language (via ResolveOrDefault(string)) instead of falling back to the bundled English hyphenator — which would otherwise hyphenate any embedded Latin-script runs in a document tagged for such a language. Same primary-subtag normalization + validation as Register(string, string, string?).

public static void RegisterNoHyphenation(string language)

Parameters

language string

TryHyphenate(string, ReadOnlySpan<char>, out int[], int, int)

Find the hyphenation break positions for word in language. A break position k means "a soft break may go between word[k-1] and word[k]". Returns false (and an empty array) when the language isn't registered.

public static bool TryHyphenate(string language, ReadOnlySpan<char> word, out int[] breaks, int leftMin = 2, int rightMin = 3)

Parameters

language string

word ReadOnlySpan<char>

breaks int[]

leftMin int

rightMin int

Returns

bool