Previously, the ordering_incompatible check was performed unconditionally, causing errors even when the --key option was used, where such incompatibilities might not apply. This change adds a condition to skip the check if --key is present, ensuring correct behavior for key-based sorting.
Update the expected output for the multiple decimals numeric sort test to reflect the proper ascending order. The values "576,446.88800000" and "576,446.890" were misplaced and have been repositioned to their correct locations in the sorted sequence, ensuring the test accurately validates the sorting logic. The debug fixture was updated accordingly.
The assignment of NumInfo::parse result was reformatted by splitting it across two lines to enhance code readability and adhere to line length guidelines.
- Update detect_numeric_locale to check for C locale (ASCII encoding and "und" locale)
- In C locale, set thousands_sep to None to avoid incorrect grouping separators
- Adjust test expectations to match new sorting behavior for numeric fields in C locale
- Ignore thousands separators in debug annotations to match GNU output
- Simplify NumInfo parsing by removing redundant thousands separator logic
- Enhance detection of numeric locale settings to handle multibyte separators like NBSP correctly, maintaining single-byte behavior for compatibility with upstream GNU coreutils
Use struct literal initialization instead of creating a mutable default and assigning fields, improving code conciseness and readability without changing functionality.
Implement NumericLocaleSettings to handle thousands separators and decimal points based on locale. Update tokenization logic to accommodate blank thousands separators for numeric and human-numeric modes, improving parsing of locale-specific numbers. Also refactor numeric locale detection for safety/readability and clean up related initialization/spell-checker ignore.
Use array slice for trim_end_matches and String::len for length check to improve readability and efficiency in test_human_numeric_blank_thousands_sep_locale.
Add a new test function `test_human_numeric_blank_thousands_sep_locale` to verify that the sort utility correctly handles human-readable numeric sorting when the locale's thousands separator is a blank space (e.g., in sv_SE.UTF-8 or sv_SE). This ensures proper behavior of the `-h` flag with key-based sorting in such locales, preventing potential sorting errors with space-separated numeric strings.
Implement NumericLocaleSettings to handle thousands separators and decimal points based on locale. Update tokenization logic to accommodate blank thousands separators for numeric and human-numeric modes, improving parsing of locale-specific numbers. Also refactor numeric locale detection for safety/readability and clean up related initialization/spell-checker ignore.
Daniel Hofstetter
Sylvestre Ledru
mattsu
Chris Dryden
oech3
renovate[bot]
ChiamakaUI