v0.3

Release Notes — TinyPreprocessor v0.3

Breaking changes

• The library is now generic over content representation (TContent) instead of a symbol type (TSymbol + ReadOnlyMemory<TSymbol>).
• Preprocessor<…> now requires an IContentModel<TContent> instance to define:
  • how content length is measured
  • how content is sliced
  • what a source-map “offset” means (“content units”)
• Public interfaces updated accordingly:
  • IResource<TContent> / Resource<TContent> now store TContent Content
  • IResourceResolver<TContent> and ResourceResolutionResult<TContent>
  • IDirectiveParser<TContent, TDirective> parses TContent directly
  • IMergeStrategy<TContent, TDirective, TContext>, ResolvedResource<TContent, TDirective>, MergeContext<TContent, TDirective>
  • PreprocessResult<TContent> now returns TContent Content

New

• Added IContentModel<TContent> abstraction.
• Added a text implementation for the default text pipeline (ReadOnlyMemory<char>).

Text pipeline

• ConcatenatingMergeStrategy updated to the new merge contracts and continues to operate as the default text merge strategy using TContent = ReadOnlyMemory<char>.

Source maps

• Source mapping remains offset-based; documentation now reflects the current implementation.
• Offsets are now explicitly “content units” as defined by the supplied IContentModel<TContent>.

Documentation

• README and architecture docs updated to match the new TContent + IContentModel<TContent> APIs and examples.

Testing

• Test suite updated to the new API and remains green (109 tests passing).
• Existing xUnit analyzer warnings remain (no behavioral change implied).

Upgrade guidance (high level)

• Replace TSymbol usages with TContent in your implementations.
• Provide an IContentModel<TContent> when constructing Preprocessor<…>.
• If you’re using the text pipeline, use TContent = ReadOnlyMemory<char> and ReadOnlyMemoryCharContentModel.