Enhance YAML parser with flow collection preprocessing and tag handling

[rjrodger]

Summary

This PR significantly improves YAML parsing capabilities by adding comprehensive flow collection preprocessing and proper TAG directive handling. The changes enable the parser to handle YAML-specific features that Jsonic's core parser doesn't natively support, while also fixing edge cases in block scalar and multiline key handling.

Key Changes

Flow Collection Preprocessing

• Added preprocessFlowCollections() function to transform flow collection syntax into Jsonic-compatible format before parsing
• Handles implicit null-valued keys in flow mappings: {a, b: c} → {a: ~, b: c}
• Strips comments between keys and colons in flow context
• Folds multiline quoted scalars into single lines
• Supports explicit key indicators (?) inside flow collections
• Converts [? key : val] syntax to [{key: val}] for proper parsing

TAG Directive Support

• Added tagHandles map to track %TAG directive mappings
• Parses and stores TAG directives before document processing
• Skips built-in type conversion (int, float, bool, null) when !! is redefined by a custom TAG directive
• Allows users to override YAML's default type tags with custom handlers

Block Scalar Key Handling

• Added support for block scalar indicators (| and >) in explicit keys
• Properly handles chomping indicators (+, -) and explicit indentation
• Correctly processes folded vs. literal block scalars in key position
• Tracks extra rows consumed for accurate position tracking

Alias Resolution Improvements

• Resolves aliases immediately when anchors are available, preventing loss of markers through Jsonic's rule processing
• Defers resolution only when anchor hasn't been seen yet
• Properly handles alias values as map keys

Tag Processing Enhancements

• Handles standalone tags (followed by newline) by consuming the newline and leading spaces
• Prevents spurious indent tokens after standalone tags
• Correctly identifies tag boundaries when colons appear in tag names

Sequence Marker Detection

• Improved sequence marker (- ) detection to only treat as new entry when indent matches enclosing sequence level
• Prevents false positives in continuation lines

Position Tracking Fixes

• Fixed column index calculation for explicit keys with values
• Properly tracks row increments for multiline keys and block scalars
• Corrects continuation line indent calculation for map values

Test Suite Updates

• Removed 14 test cases from SKIP list that now pass with these improvements
• Tests cover flow collections, block scalars, tags, aliases, and complex nesting scenarios

https://claude.ai/code/session_01H3rUS9E1u5eXZrzyYiBPB3