sort: gnu coreutils compatibility (sort float.sh) by mattsu2020 · Pull Request #9839 · uutils/coreutils

mattsu2020 · 2025-12-25T12:16:50Z

We have made corrections to ensure the GNU coreutils tests pass.

sort-float.sh

github-actions · 2025-12-25T12:45:27Z

GNU testsuite comparison:

Congrats! The gnu test tests/sort/sort-float is no longer failing!

sylvestre · 2025-12-25T13:28:59Z

Could you please add a test to make sure we don't regress in the future?
Thanks

sylvestre · 2025-12-25T13:30:01Z

Did you run some benchmark on this btw?

github-actions · 2025-12-26T00:42:36Z

GNU testsuite comparison:

Congrats! The gnu test tests/sort/sort-float is no longer failing!

github-actions · 2025-12-29T02:21:00Z

GNU testsuite comparison:

Congrats! The gnu test tests/sort/sort-float is no longer failing!

Add locale-aware handling for decimal separators (comma vs. period) in general numeric sorting mode. This allows proper sorting of numeric fields in locales where comma is the standard decimal separator, such as many European countries. Changes include: - Enabling the "i18n-decimal" feature in uucore dependencies - Introducing locale_decimal_pt() to retrieve the locale-specific decimal separator - Updating get_leading_gen() and general_bd_parse() to accept and use the decimal point parameter - Converting input to standard period notation for parsing when locale uses comma This fixes sorting inaccuracies for international users with comma-based decimal notation.

…g Option::then Refactored the `general_bd_parse` function to replace manual mutable vector initialization and conditional assignment with `Option::then` and `as_deref`, making the code more concise and idiomatic while maintaining the same logic for normalizing decimal points.

Add `effective_decimal_pt` function to intelligently select the decimal point based on input content, ensuring correct sorting when locale uses comma but input uses period as decimal separator. This fixes parsing issues in general numeric mode by prioritizing input-based detection over strict locale adherence.

Updated various crates in fuzz/Cargo.lock: windows-sys to 0.61.2, bumpalo to 3.19.1, cc to 1.2.50, crc to 3.3.0, crc-fast to 1.9.0, icu_properties and data to 2.1.2, jiff to 0.2.17. Added fixed_decimal, icu_decimal, and icu_decimal_data for improved functionality and compatibility.

…ntains method Use input.contains() instead of input.iter().any() for checking presence of comma and dot bytes, improving readability and reducing boilerplate.

Split the uucore dependency features into multiple lines to improve code formatting and maintain consistency with other multi-line dependency declarations.

Add test to verify that sort's general numeric (-g) option correctly handles decimal separators based on locale settings, ensuring proper ordering of numbers like "1,9" vs "1,10" in French locale (comma as separator).

ChrisDryden · 2026-01-07T00:10:17Z

src/uu/sort/src/sort.rs

+    }
+}
+
+fn effective_decimal_pt(input: &[u8], locale_decimal: u8) -> u8 {


Why is this function used? Shouldn't it always just default to the locale?

ChrisDryden · 2026-01-07T00:16:07Z

src/uu/sort/src/sort.rs

+                    let locale_decimal = locale_decimal_pt();
+                    let decimal_pt = effective_decimal_pt(initial_selection, locale_decimal);


I think this can just be:

let decimal_pt = locale_decimal_pt();

An example of where this causes issues is:

echo -e "1.9\n1.1" | LC_ALL=fr_FR.utf8 /usr/bin/sort -g --stable # GNU output: 1.9, 1.1 (both = 1, stable) # PR output: 1.1, 1.9 (WRONG - treats as floats) echo -e "2,5\n1.9\n1,1" | LC_ALL=fr_FR.utf8 /usr/bin/sort -g # GNU output: 1.9, 1,1, 2,5 (values: 1, 1.1, 2.5) # PR output: 1,1, 1.9, 2,5 (WRONG - inconsistent decimal handling per line) echo -e "1.5e10\n2" | LC_ALL=fr_FR.utf8 /usr/bin/sort -g # GNU output: 1.5e10, 2 (1 < 2) # PR output: 2, 1.5e10 (WRONG - treats as 15000000000 > 2)

Would be good to add these cases to the GNU test suite, will follow up with that

ChrisDryden · 2026-01-07T00:16:41Z

src/uu/sort/src/sort.rs

            // Parse this number as BigDecimal, as this is the requirement for general numeric sorting.
-            Selection::AsBigDecimal(general_bd_parse(&range_str[get_leading_gen(range_str)]))
+            let locale_decimal = locale_decimal_pt();
+            let decimal_pt = effective_decimal_pt(range_str, locale_decimal);


Same here for the comment above, you can take that entire helper function out

…e_decimal_pt Remove the effective_decimal_pt function and update its usages in Line and FieldSelector to directly call locale_decimal_pt, simplifying the code without altering functionality. This refactor eliminates unnecessary logic for determining decimal points in general numeric sort mode.

github-actions · 2026-01-07T07:12:06Z

GNU testsuite comparison:

Skipping an intermittent issue tests/timeout/timeout (passes in this run but fails in the 'main' branch)
Congrats! The gnu test tests/sort/sort-float is no longer failing!

sylvestre · 2026-01-07T07:26:52Z

Did you run some benchmark on this btw?

ping?

mattsu2020 · 2026-01-07T09:27:33Z

Did you run some benchmark on this btw?

ping?

Since regression is occurring, I will investigate.

github-actions · 2026-01-07T11:32:02Z

GNU testsuite comparison:

Skip an intermittent issue tests/timeout/timeout (fails in this run but passes in the 'main' branch)
Congrats! The gnu test tests/sort/sort-float is no longer failing!

codspeed-hq · 2026-01-07T11:38:42Z

Merging this PR will not alter performance

✅ 141 untouched benchmarks
⏩ 38 skipped benchmarks¹

_{Comparing mattsu2020:sort_sort-float.sh (2f14b2d) with main (574f6ba)}

38 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

github-actions · 2026-01-08T11:38:46Z

GNU testsuite comparison:

Congrats! The gnu test tests/sort/sort-float is no longer failing!

github-actions · 2026-01-13T21:51:59Z

GNU testsuite comparison:

Congrats! The gnu test tests/sort/sort-float is no longer failing!

* feat(sort): support international decimal separators in numeric sorting M

ChrisDryden mentioned this pull request Dec 26, 2025

GNU coreutils 9.9: Detailed Test report 12/19 #9729

Closed

mattsu2020 added 7 commits December 31, 2025 20:16

refactor(sort): simplify byte checks in effective_decimal_pt using co…

81ec542

…ntains method Use input.contains() instead of input.iter().any() for checking presence of comma and dot bytes, improving readability and reducing boilerplate.

style(sort): format uucore features list in Cargo.toml for readability

3751b81

Split the uucore dependency features into multiple lines to improve code formatting and maintain consistency with other multi-line dependency declarations.

test: add locale-aware decimal separator test for sort -g

8da07b8

Add test to verify that sort's general numeric (-g) option correctly handles decimal separators based on locale settings, ensuring proper ordering of numbers like "1,9" vs "1,10" in French locale (comma as separator).

sylvestre force-pushed the sort_sort-float.sh branch from d958bc3 to 8da07b8 Compare December 31, 2025 19:16

ChrisDryden reviewed Jan 7, 2026

View reviewed changes

mattsu2020 mentioned this pull request Jan 7, 2026

bench(sort): add general numeric benchmark #10101

Merged

Merge branch 'main' into sort_sort-float.sh

2366786

Merge branch 'main' into sort_sort-float.sh

38c7316

Merge branch 'main' into sort_sort-float.sh

2f14b2d

sylvestre merged commit 6b49ff9 into uutils:main Jan 14, 2026
132 checks passed

mattsu2020 deleted the sort_sort-float.sh branch January 14, 2026 07:03

mattsu2020 added a commit to mattsu2020/coreutils that referenced this pull request Jan 23, 2026

sort: gnu coreutils compatibility (sort float.sh) (uutils#9839)

ab98b05

* feat(sort): support international decimal separators in numeric sorting M

moonfruit mentioned this pull request Feb 3, 2026

uutils-selected 0.6.0 moonfruit/homebrew-tap#453

Closed

		let locale_decimal = locale_decimal_pt();
		let decimal_pt = effective_decimal_pt(initial_selection, locale_decimal);

Uh oh!

Conversation

mattsu2020 commented Dec 25, 2025

Uh oh!

github-actions bot commented Dec 25, 2025

Uh oh!

sylvestre commented Dec 25, 2025

Uh oh!

sylvestre commented Dec 25, 2025

Uh oh!

github-actions bot commented Dec 26, 2025

Uh oh!

github-actions bot commented Dec 29, 2025

Uh oh!

ChrisDryden Jan 7, 2026

Choose a reason for hiding this comment

Uh oh!

mattsu2020 Jan 7, 2026

Choose a reason for hiding this comment

Uh oh!

ChrisDryden Jan 7, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

ChrisDryden Jan 7, 2026

Choose a reason for hiding this comment

Uh oh!

mattsu2020 Jan 7, 2026

Choose a reason for hiding this comment

Uh oh!

ChrisDryden Jan 7, 2026

Choose a reason for hiding this comment

Uh oh!

mattsu2020 Jan 7, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions bot commented Jan 7, 2026

Uh oh!

sylvestre commented Jan 7, 2026

Uh oh!

mattsu2020 commented Jan 7, 2026

Uh oh!

github-actions bot commented Jan 7, 2026

Uh oh!

codspeed-hq bot commented Jan 7, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Merging this PR will not alter performance

Footnotes

Uh oh!

github-actions bot commented Jan 8, 2026

Uh oh!

github-actions bot commented Jan 13, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

ChrisDryden Jan 7, 2026 •

edited

Loading

codspeed-hq bot commented Jan 7, 2026 •

edited

Loading