* feat: add compact float formatting for half and bfloat16 in od

Implement trim_float_repr() to remove trailing zeros from float strings while preserving signs and exponents, and pad_float_repr() to align trimmed floats to fixed width. Update format_item_f16() and format_item_bf16() to produce compact output matching GNU od. Add regression tests for float16 and bfloat16 compact printing.

* refactor(od): format multiline format! in format_item_bf16 for readability

Reformat the format! macro call in the format_item_bf16 function in prn_float.rs
to span multiple lines, improving code readability without changing functionality.

* fix(od): preserve canonical precision for f16/bf16 float formats

Remove trimming of trailing zeros from f16 and bf16 float representations
in od output to maintain original precision and align behavior with
f32/f64 formatters, ensuring stable output across platforms. Update
corresponding tests to reflect the change in expected output.

* refactor(od): simplify float padding format and update tests

- Remove redundant `width = width` parameter from `format!` macro in `pad_float_repr`
- Add "bfloat" to spell-checker ignore list for better test coverage on bf16 format

* feat(od): trim trailing zeros in float outputs for GNU compatibility

Add `trim_trailing_zeros` function to remove trailing zeros and redundant decimal points from formatted floats, ensuring compact output matching GNU od for f16 and bf16 types. Update `format_item_f16` and `format_item_bf16` to apply trimming before padding.

* fix: preserve trailing zeros in F16 and BF16 float formats to match GNU od output

Remove the `trim_trailing_zeros` function and update `format_item_f16` and `format_item_bf16` to keep the raw formatted strings without trimming trailing zeros. This ensures consistent column widths and aligns with GNU od behavior for 16-bit float representations, preventing misalignment in output tables.

* refactor(od/prn_float): combine multiline format! into single line in format_item_f16

The format! macro call in format_item_f16 was split across multiple lines with newlines. This change consolidates it into a single line for improved code readability and consistency with similar patterns in the file, without altering the function's output or logic.

* feat(od): trim trailing zeros in half-precision and bfloat16 float outputs

- Add `trim_float_repr` function to remove unnecessary trailing zeros and padding from normalized float strings, leaving exponents unchanged.
- Update `format_item_f16` and `format_item_bf16` to apply trimming while maintaining column alignment via re-padding.
- Update test expectations to reflect the more compact float representations (e.g., "1" instead of "1.0000000").

* refactor: simplify float trimming condition in prn_float.rs

Replace `if let Some(_) = s.find('.')` with `s.find('.').is_some()` in the `trim_float_repr` function to improve code clarity and idiomatic Rust usage while maintaining the same logic for checking decimal presence=black.

* Update src/uu/od/src/prn_float.rs

Co-authored-by: Daniel Hofstetter <[email protected]>

---------

Co-authored-by: Daniel Hofstetter <[email protected]>