feat: add token source position tracking to lexer

[label-hook]

Summary

Adds source position tracking to sqlparse tokens by extending the token format from 2-tuples (ttype, value) to 3-tuples (ttype, value, position) where position is a tuple of (start, end) character indices. This enables syntax highlighting and other position-aware use cases while maintaining backwards compatibility.

Changes

• Modified lexer (sqlparse/lexer.py): Enhanced tokenization to track character positions and emit tokens with position information
• Updated token filters (sqlparse/filters/tokens.py): Modified filters to handle 3-tuple tokens while preserving existing functionality
• Enhanced filter stack (sqlparse/engine/filter_stack.py): Updated to properly process position-aware token streams
• Updated SQL token handling (sqlparse/sql.py): Modified token processing to work with extended token format

The implementation is designed to be minimally invasive - the main API surface area documented in the official docs remains unchanged. Only direct lexer usage is affected by the token format extension.

Testing

• All existing tests pass, confirming backwards compatibility
• Added specific tests for position tracking functionality
• Verified that tokens contain correct start/end character positions
• Confirmed that syntax highlighting and other position-aware use cases work as expected

Closes #24

---

Closes #24