Make ls pass the GNU ls-misc.pl suite by karanabe · Pull Request #9262 · uutils/coreutils

karanabe · 2025-11-13T15:26:12Z

Align ls’s symlink coloring with GNU semantics—or=, mi= and ln=target now behave exactly like /bin/ls, including emitting resets for blank or= entries—and add Rust integration tests that mirror the failing sl-dangle3..9 cases so regressions are caught locally.

#9127

Ensure StyleManager honors `or=`, `mi=` and `ln=target` the same as GNU ls when coloring dangling symlinks. Reset sequences are still printed for blank `or=` entries, and targets fall back to missing-file colors.

Add seven ls tests mirroring GNU tests/ls/ls-misc.pl sl-dangle3..9 to validate combinations of `ln=`, `or=` and `mi=` settings. The tests verify both file-level output and directory listings, including the `\x1b[m` reset requirement when `or=:` is set. Adjust existing Rust tests to GNU output. Existing color tests were updated to match GNU output.

karanabe · 2025-11-13T15:34:42Z

Existing Rust color tests were updated so their expected output now matches GNU ls.

github-actions · 2025-11-13T15:59:57Z

GNU testsuite comparison:

GNU test failed: tests/ls/color-dtype-dir. tests/ls/color-dtype-dir is passing on 'main'. Maybe you have to rebase?
GNU test failed: tests/ls/color-norm. tests/ls/color-norm is passing on 'main'. Maybe you have to rebase?
GNU test failed: tests/ls/multihardlink. tests/ls/multihardlink is passing on 'main'. Maybe you have to rebase?
GNU test failed: tests/ls/stat-free-color. tests/ls/stat-free-color is passing on 'main'. Maybe you have to rebase?
Skip an intermittent issue tests/tail/overlay-headers (fails in this run but passes in the 'main' branch)
Congrats! The gnu test tests/ls/ls-misc is no longer failing!

github-actions · 2025-11-13T17:43:07Z

GNU testsuite comparison:

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

Colors: make StyleManager rely on lscolors’ explicit-style flags, avoid unnecessary metadata/stat probes, and ensure ln=target, or=, mi= honor GNU coloring semantics. Tests: update test_ls_color_norm expectations so the fi= and fi=0 cases now match GNU ls output, ensuring the new color handling is verified.

github-actions · 2025-11-13T18:12:03Z

GNU testsuite comparison:

Skip an intermittent issue tests/tail/overlay-headers (fails in this run but passes in the 'main' branch)
Congrats! The gnu test tests/ls/ls-misc is no longer failing!

- factor ANSI escape literals into named constants - document why we bypass LsColors fallbacks when applying raw SGR codes

github-actions · 2025-11-14T13:08:04Z

GNU testsuite comparison:

Skip an intermittent issue tests/tail/overlay-headers (fails in this run but passes in the 'main' branch)
Congrats! The gnu test tests/ls/ls-misc is no longer failing!

codspeed-hq · 2025-11-14T16:31:22Z

CodSpeed Performance Report

Merging this PR will not alter performance

_{Comparing karanabe:test/fix-ls-mics (a9b5ab5) with main (b804ca4)}

Summary

✅ 140 untouched benchmarks
⏩ 37 skipped benchmarks¹

37 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 · 2025-11-14T17:08:24Z

GNU testsuite comparison:

Skip an intermittent issue tests/tail/overlay-headers (fails in this run but passes in the 'main' branch)
Congrats! The gnu test tests/ls/ls-misc is no longer failing!

karanabe · 2025-11-14T17:38:07Z

The CI failure is caused by a different issue (refer to #9276), not by this change.
CICD / Test all features separately (ubuntu-latest, feat_os_unix) (pull_request)

sylvestre · 2025-11-14T22:11:03Z

I was wondering if some of these changes should not in the crate lscolor. Wdyt? Thanks

karanabe · 2025-11-15T03:30:59Z

We already lean on lscolors for the generic bits (parsing LS_COLORS, providing Indicator styles, etc.). The code here lives on top because it’s tightly coupled to StyleManager state (indicator_codes, ln=target handling, PathData-based stat avoidance) and to the GNU-specific behavior we emulate. lscolors’ public API doesn’t expose those knobs, so moving this logic there would either require a significant API expansion or reintroduce the same context we already have on our side. For now it’s cleaner to keep the GNU-specific tailoring in ls.rs/colors.rs and continue reusing lscolors for the common functionality.

In the future, if more consumers of lscolors run into the same GNU-specific needs (e.g., ln=target, empty or= semantics, stat avoidance hooks), we can revisit the API with the maintainers and see whether it makes sense to expose those knobs upstream. Until then, keeping this logic local lets us iterate quickly while still benefitting from lscolors for the generic parts.

sylvestre · 2025-11-15T08:56:06Z

@karanabe ok, thanks!
@sharkdp in case there is anything you would like to cherrypick for lscolors :)

github-actions · 2025-11-15T16:45:46Z

GNU testsuite comparison:

Congrats! The gnu test tests/ls/ls-misc is no longer failing!

karanabe · 2025-11-21T16:29:54Z

@sylvestre Thanks for the review! If everything looks good, could you proceed with the merge?

sylvestre · 2025-12-26T23:37:19Z

+    style_manager.apply_style_based_on_metadata(path, md_option.as_ref(), name, wrap)
+}
+
+fn parse_indicator_codes() -> (HashMap<Indicator, String>, bool) {


maybe it could benefit from better error handling for malformed LS_COLOR

Added GNU-like LS_COLORS validation with warnings and color disabled on parse errors.

sylvestre · 2025-12-26T23:39:07Z

+fn escape_dir_name_with_locale(name: &OsStr, config: &Config) -> OsString {
+    if let Some(locale) = config.locale_quoting {
+        locale_quote(name, locale)
+    } else {
+        locale_aware_escape_dir_name(name, config.quoting_style)
+    }
+}
+
+fn escape_name_with_locale(name: &OsStr, config: &Config) -> OsString {
+    if let Some(locale) = config.locale_quoting {
+        locale_quote(name, locale)
+    } else {
+        locale_aware_escape_name(name, config.quoting_style)
+    }
+}


i think it could be dedup a bit

Deduplicated locale escaping by introducing a shared helper.

sylvestre · 2025-12-26T23:41:05Z

-                style_code.push_str(self.reset(!self.initial_reset_is_done));
-                style_code.push_str(&self.get_style_code(new_style));
+        if let Some(path) = path {
+            // Fast-path: apply LS_COLORS raw SGR codes verbatim,


maybe move this block into a function to decrease complexity ?

Extracted raw LS_COLORS application into helpers to reduce complexity.

sylvestre · 2025-12-26T23:41:10Z

-            force_suffix_reset = true;
+
+        if !applied_raw_code {
+            if let Some(new_style) = new_style {


maybe move this block into a function to decrease complexity ?

Moved fallback style application into a helper to simplify the flow.

karanabe · 2025-12-29T23:41:06Z

Thanks for the review! I’ll address the review comments.

sylvestre · 2026-01-06T17:45:18Z

@karanabe ping ?

Add GNU-like LS_COLORS validation and disable color output on parse errors with warnings. Refactor raw style application and fallback handling to reduce complexity and document empty entries. De-duplicate locale escaping and add warning strings for LS_COLORS errors.

karanabe · 2026-01-07T13:01:13Z

Thanks for the ping. I’ve made a small update to this PR.

This update adds GNU-like LS_COLORS validation and refactors style application to reduce complexity.
Locale escaping was de-duplicated.

github-actions · 2026-01-07T13:08:32Z

GNU testsuite comparison:

Congrats! The gnu test tests/ls/ls-misc is no longer failing!

sylvestre · 2026-01-07T13:15:33Z

+    validate_ls_colors(&ls_colors)
+}
+
+fn validate_ls_colors(ls_colors: &str) -> Result<(), LsColorsParseError> {


i think this function could benefit from some comments

Added a short comment describing the GNU-like LS_COLORS validation intent.

sylvestre · 2026-01-07T13:15:47Z

+    }
+
+    let mut state = State::Ground;
+    loop {


Added a short comment explaining that the parser handles GNU-compatible escapes.

sylvestre · 2026-01-07T13:17:48Z

+        if raw.is_empty() {
+            Some(RawIndicatorStyle::Empty)
+        } else {
+            Some(RawIndicatorStyle::Code(raw.clone()))


please remove the clone()

Removed the clone by switching to an Indicator-based path and re-fetching the raw code.

github-actions · 2026-01-07T14:53:49Z

GNU testsuite comparison:

Congrats! The gnu test tests/ls/ls-misc is no longer failing!

fix(ls): match GNU dangling color semantics

a764ba0

Ensure StyleManager honors `or=`, `mi=` and `ln=target` the same as GNU ls when coloring dangling symlinks. Reset sequences are still printed for blank `or=` entries, and targets fall back to missing-file colors.

karanabe force-pushed the test/fix-ls-mics branch from 81fd4d3 to 2dc62c3 Compare November 13, 2025 15:32

karanabe force-pushed the test/fix-ls-mics branch from 2dc62c3 to 90e54fb Compare November 13, 2025 15:33

karanabe marked this pull request as draft November 13, 2025 16:09

karanabe force-pushed the test/fix-ls-mics branch from c0088ea to 480b198 Compare November 13, 2025 17:45

karanabe force-pushed the test/fix-ls-mics branch from 480b198 to 692c1a1 Compare November 13, 2025 17:50

sylvestre reviewed Nov 13, 2025

View reviewed changes

Comment thread src/uu/ls/src/colors.rs Outdated

sylvestre reviewed Nov 13, 2025

View reviewed changes

Comment thread src/uu/ls/src/colors.rs Outdated

sylvestre reviewed Nov 13, 2025

View reviewed changes

Comment thread src/uu/ls/src/ls.rs Outdated

karanabe added 2 commits November 14, 2025 21:23

lint(ls): clarify ANSI escape handling

c365b44

- factor ANSI escape literals into named constants - document why we bypass LsColors fallbacks when applying raw SGR codes

refactor(ls): structure quoting style match

af8afe0

karanabe force-pushed the test/fix-ls-mics branch 3 times, most recently from 6164a04 to af8afe0 Compare November 14, 2025 16:19

Merge branch 'main' into test/fix-ls-mics

5c0cd48

karanabe marked this pull request as ready for review November 14, 2025 17:38

karanabe requested a review from sylvestre November 14, 2025 17:38

Merge branch 'main' into test/fix-ls-mics

dab76b9

sharkdp mentioned this pull request Nov 15, 2025

Consider adopting downstream fixes in coreutils sharkdp/lscolors#93

Open

ChrisDryden mentioned this pull request Dec 24, 2025

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

Closed

sylvestre reviewed Dec 26, 2025

View reviewed changes

Merge branch 'main' into test/fix-ls-mics

3a97e7f

sylvestre reviewed Jan 7, 2026

View reviewed changes

karanabe added 2 commits January 7, 2026 23:04

Refactor LS_COLORS handling and comments

036a13d

Merge branch 'main' into test/fix-ls-mics

a9b5ab5

sylvestre merged commit 0b0968a into uutils:main Jan 8, 2026
131 checks passed

karanabe deleted the test/fix-ls-mics branch January 17, 2026 07:28

moonfruit mentioned this pull request Feb 3, 2026

uutils-selected 0.6.0 moonfruit/homebrew-tap#453

Closed

Uh oh!

Uh oh!

Conversation

karanabe commented Nov 13, 2025

Uh oh!

karanabe commented Nov 13, 2025

Uh oh!

github-actions Bot commented Nov 13, 2025

Uh oh!

github-actions Bot commented Nov 13, 2025

Uh oh!

github-actions Bot commented Nov 13, 2025

Uh oh!

Uh oh!

Uh oh!

Uh oh!

github-actions Bot commented Nov 14, 2025

Uh oh!

codspeed-hq Bot commented Nov 14, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Merging this PR will not alter performance

Summary

Footnotes

Uh oh!

github-actions Bot commented Nov 14, 2025

Uh oh!

karanabe commented Nov 14, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

sylvestre commented Nov 14, 2025

Uh oh!

karanabe commented Nov 15, 2025

Uh oh!

sylvestre commented Nov 15, 2025

Uh oh!

github-actions Bot commented Nov 15, 2025

Uh oh!

karanabe commented Nov 21, 2025

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

karanabe commented Dec 29, 2025

Uh oh!

sylvestre commented Jan 6, 2026

Uh oh!

karanabe commented Jan 7, 2026

Uh oh!

github-actions Bot commented Jan 7, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

github-actions Bot commented Jan 7, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

codspeed-hq Bot commented Nov 14, 2025 •

edited

Loading

karanabe commented Nov 14, 2025 •

edited

Loading