Provide source locations for parsed specification item IDs

[kaklakariada]

Problem

OFT parsing currently exposes only an item-level file and line location. Consumers need precise source locations for the individual IDs within a parsed specification item, without parsing the source text again.

Required information

For every complete parsed SpecificationItem, make the following information available:

• The location/range of its declared ID.
• The location/range of every covered ID.
• The location/range of every dependency ID.
• For each located ID:
  • whole-ID range;
  • artifact-type range;
  • name range;
  • revision range.
• The semantic SpecificationItemId associated with each occurrence.
• Distinct locations for repeated occurrences of the same semantic ID.
• Locations for source constructs whose resulting IDs are partly generated by OFT, including the ranges of every component actually present in source.
• A documented coordinate convention suitable for editor protocols, including line/column indexing, character encoding, and whether end positions are inclusive or exclusive.

Use cases

• Semantic highlighting of individual ID components.
• textDocument/semanticTokens/full.
• textDocument/semanticTokens/range.
• Future completion of ID artifact types, names, and revisions.
• Navigation and reference features without duplicate OFT parsing.

Acceptance criteria

• Information is available for every OFT importer.
• It is produced from OFT’s parser rather than reconstructed by consumers.
• Tests cover declarations, covered IDs, dependency IDs, repeated IDs, multiple IDs, multiline constructs, and abbreviated/generated source syntax.