…fy function signatures

Introduce a new FoldContext struct to group related fields (spaces, width, mode, writer, output, col_count, last_space) into a single context object. This refactoring reduces parameter passing in emit_output and process_utf8_line functions, improving code readability and maintainability without altering the core folding logic.

Add "rposition" to the cspell jargon dictionary to prevent spell check errors for this technical term.

…output

Break the inline closure into a multi-line block for better code clarity and maintainability.

The condition for updating the last space index was changed from `idx + 1 <= consume` to `idx < consume` to fix an off-by-one error, ensuring proper handling of spaces when consuming characters during line folding.

…andling

Refactor the `process_ascii_line` function to use a while loop with pattern matching instead of a for loop, improving efficiency and clarity. Introduce `push_ascii_segment` to handle contiguous printable character sequences, ensuring accurate column counting and whitespace tracking in both columns and characters modes. This addresses potential issues with control character processing and width calculations.

…fy function signatures

Introduce a new FoldContext struct to group related fields (spaces, width, mode, writer, output, col_count, last_space) into a single context object. This refactoring reduces parameter passing in emit_output and process_utf8_line functions, improving code readability and maintainability without altering the core folding logic.

Add process_ascii_line function to handle ASCII bytes efficiently, avoiding UTF-8 overhead for ASCII input. Update emit_output to properly manage output buffer remainder and track last space position for better folding logic. Modify process_utf8_line to delegate ASCII lines to the new function.

Add "rposition" to the cspell jargon dictionary to prevent spell check errors for this technical term.

- Replace fold/writeln! with loop/push_str in benchmarks for faster string building
- Add append_usize helper to avoid allocations in benchmark data generation
- Refactor emit_output to use drain instead of split_off for better performance
- Update last_space calculation to handle index adjustments more efficiently

These changes improve performance in the fold utility's benchmarks and core logic by reducing allocations and optimizing string operations.

…output

Break the inline closure into a multi-line block for better code clarity and maintainability.

The condition for updating the last space index was changed from `idx + 1 <= consume` to `idx < consume` to fix an off-by-one error, ensuring proper handling of spaces when consuming characters during line folding.

…andling

Refactor the `process_ascii_line` function to use a while loop with pattern matching instead of a for loop, improving efficiency and clarity. Introduce `push_ascii_segment` to handle contiguous printable character sequences, ensuring accurate column counting and whitespace tracking in both columns and characters modes. This addresses potential issues with control character processing and width calculations.