v0.4.3

Release Notes: v0.4.3

• Fixed line/column mapping inaccuracies reported by downstream users when source text contains complex newline patterns and trivia (e.g., comments).
• Improved line boundary detection to ensure correct handling of:
  • LF / CRLF / mixed line endings
  • leading blank lines and consecutive blank lines
  • trailing newline producing a final empty line
  • multi-line /* ... */ comments containing embedded newlines
• Added extensive unit test coverage for SyntaxTreeLineColumnMapper and SyntaxTreeLineBoundaryResolver, including strict contract checks (offsets are unique, strictly increasing, and within range).

v0.4.2

Release Notes v0.4.2

Improvements

• Improved source mapping for preprocessed output so merged offsets can be traced back to the correct original resource when using imports (including multiple imports and nested imports).

Testing / Reliability

• Added an end-to-end integration test that validates source-map resolution across multiple imported resources.

Technical Notes

• Source-map segments are now generated during the merge pipeline to avoid redundant tree walks.
• Mapping boundaries account for SyntaxEditor.Replace(...) trivia transfer by splicing on the import directive’s token text span (so preserved trivia remains attributed to the importing file).

v0.4.1

Release Notes — Diagnostics Line/Column + C#-Style Output

• Added line/column mapping for SyntaxTree diagnostics using IContentBoundaryResolver<SyntaxTree, LineBoundary> so locations can be reported as 1-based line,column (and spans as line,col,endLine,endCol).
• Introduced a C#-compiler-style diagnostic formatter: resource(line,column): severity CODE: message.
• Improved merge-time diagnostics to prefer human-friendly line:column output when available (instead of raw offset ranges).
• Added/updated unit tests covering line boundary mapping and diagnostic formatting.

Compatibility Notes

• Diagnostic text formatting may change compared to previous versions (locations are now line/column oriented and more compiler-like).

v0.4

Release Notes — TinyAst.Preprocessor 0.4.0

Highlights

• Upgraded to TinyPreprocessor v0.4 and aligned merge behavior with v0.4 semantics.
• Added line-boundary support for SyntaxTree source-map boundary resolution.
• Expanded test coverage for canonical ResourceId scenarios and redundant merge work prevention.

Breaking / Behavioral changes

• Merge identity is now resolver-authoritative:
  • ResourceId is treated as opaque during merge (not required to be path-like).
  • Merge uses the resolver-produced resolved-id mapping (MergeContext.ResolvedReferences) per directive occurrence.
  • Merge no longer re-derives dependency IDs from raw reference strings (path heuristics).

New features

• Boundary resolution support (Line boundaries):
  • Implemented IContentBoundaryResolver<SyntaxTree, LineBoundary> and a provider wired into the pipeline.
  • Enables boundary-aware source-map lookups (e.g., line index resolution) via TinyPreprocessor’s boundary extensions.

Improvements

• No redundant merge work (validated by tests):
  • Shared dependencies and include-of-include graphs are processed exactly once per ResourceId.
• Test instrumentation seam:
  • Added a protected virtual OnProcessResource(ResourceId) hook to observe per-resource processing in tests.

Documentation

• Updated README and bridge docs to reflect:
  • TinyPreprocessor v0.4 requirement
  • resolved-id mapping merge semantics
  • boundary resolution capability
• Added changelog and linked it from README.

Quality

• Added XML documentation for public/protected APIs (eliminates CS1591 warnings).

v0.1

TinyAst.Preprocessor v0.1.0 — Release Notes (First Release)

Highlights

• AST-native preprocessing on schema-bound SyntaxTree (no text re-parsing).
• Import discovery via TinyAst Query (Query.Syntax<TImportNode>()) with locations anchored to node start (Position..Position).
• Merge implemented via TinyAst SyntaxEditor (no green-node manipulation).
• Full TinyPreprocessor 0.3.0+ pipeline support: dependency resolution, diagnostics, and source mapping.

New APIs

• ImportDirective + ImportDirectiveModel for directive semantics and location modeling.
• ImportDirectiveParser<TImportNode> (delegate-based): configure with Func<TImportNode, string?> getReference.
• SyntaxTreeMergeStrategy<TImportNode, TContext> (delegate-based): same reference extractor used during merge.
• SyntaxTreeContentModel implementing IContentModel<SyntaxTree>.
• Test helpers: InMemorySyntaxTreeResourceStore, InMemorySyntaxTreeResourceResolver, ImportDirectiveLocationIndex.
• Convenience wiring:
  • SyntaxTreePreprocessor<TImportNode, TContext> (one-stop preprocessor setup)
  • SyntaxTreePreprocessor<TImportNode> (context defaults to object)
  • SyntaxTreeBridge<TImportNode, TContext> (parser + merge strategy bundle)

Behavior / Design Notes

• Schema binding is required (SyntaxTree.HasSchema == true) for directive discovery.
• Reference extraction is non-invasive: downstream AST nodes do not need to implement any bridge interface.
• Resolution failures can be pinned back to import locations via ImportDirectiveLocationIndex.

Diagnostics & Limits

• Cycle detection (CircularDependencyDiagnostic) and max include depth enforcement (MaxDepthExceededDiagnostic) are covered by end-to-end tests.
• Resolution failures produce ResolutionFailedDiagnostic and can carry location/resource context.

Packaging

• NuGet packaging is configured and dotnet pack produces TinyAst.Preprocessor.0.1.0.nupkg.
• Includes symbols (.snupkg) and generated XML documentation output (currently with CS1591 warnings for missing docs).

Compatibility

• Target framework: .NET 8 (net8.0)
• Depends on: TinyAst 0.11.0, TinyPreprocessor 0.3.0

Quality

• Test suite: 76 tests passing (dotnet test green), including integration coverage of the full preprocessing pipeline.