Feature proposal

Enhance the CSV parser to support fully configurable escaped fields, in alignment with the direction and edge case handling described in PHILOSOPHY.md.

Requirements:

• Configurable Escape Character:
  • Allow users to define a custom escape character for fields (default: ").
  • All parsing and serialization logic must respect this configuration.
• Escaped Field Support:
  • Escaped fields must be surrounded by the escape character.
  • Escaped fields can contain any character, including the delimiter and newline.
  • Escape characters inside an escaped field must be represented by doubling them.
  • Escaped fields must be strictly validated: leading/trailing whitespace, improper closures, or unescaped escape characters should result in errors as per the philosophy.
• Enable/Disable (De)serialization:
  • Provide an option to toggle escape field (de)serialization logic on or off. When disabled, fields should be treated as unescaped (raw) text.
• Edge Case Handling:
  • Follow all edge cases as outlined in PHILOSOPHY.md, including:
    • Trailing delimiter produces an empty field.
    • Empty escaped fields are valid.
    • Whitespace around escape characters invalidates the field.
    • Backslash escaping is not supported (fail on use).
    • Inconsistent row lengths, empty lines, and other failure cases should be handled as described.

References

• See PHILOSOPHY.md for guiding principles, supported/unsupported cases, and design rationale.

Example scenarios

• Parsing CSV rows with custom escape characters (e.g., ' instead of ").
• Ensuring invalid formats (e.g., whitespace around escape characters, missing closure) produce immediate errors.
• Optionally serializing data without any escaping.

This feature is crucial for maintaining compatibility with real-world CSV data while upholding strict data integrity and fail-fast behavior.