fix(uucore): use is_dir() instead of exists() for locale path resolution

[mattsu2020]

Summary

Fix resolve_locales_dir_from_exe_dir() in release builds to use .is_dir() instead of .exists() when checking for locale directories.

In Debian build environments where individual utility binaries (e.g. target/release/wc) exist alongside the multicall coreutils binary, .exists() incorrectly matches the binary file as a "locale directory". This causes setup_localization() to attempt filesystem-based locale loading from the wrong path, resulting in a FluentBundle that lacks utility-specific messages. The translate!() macro then returns raw Fluent keys instead of resolved translations.

Fixes the test failure reported in #11023 (comment):

assertion `left == right` failed: wc should stop after the first stdout write error:
  "wc: wc-error-failed-to-print-result: No space left on device\n"
  left: 0
  right: 1

Changes

• src/uucore/src/lib/mods/locale.rs: Changed 3 occurrences of .exists() to .is_dir() in resolve_locales_dir_from_exe_dir()

[codspeed-hq]

Merging this PR will degrade performance by 5.61%

⚠️ Different runtime environments detected

Some benchmarks with significant performance changes were compared across different runtime environments,
which may affect the accuracy of the results.

Open the report in CodSpeed to investigate

⚡ 9 improved benchmarks
❌ 6 regressed benchmarks
✅ 294 untouched benchmarks
⏩ 46 skipped benchmarks¹

⚠️ Please fix the performance issues or acknowledge them on CodSpeed.

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
⚡	Simulation	cksum_blake2b	192.7 ms	183.8 ms	+4.86%
❌	Simulation	cksum_sha256	1.7 s	1.8 s	-4.04%
❌	Simulation	cksum_crc32b	13 ms	13.7 ms	-5.61%
❌	Simulation	cksum_sha2	1.7 s	1.8 s	-4.04%
❌	Simulation	cksum_sha224	1.7 s	1.8 s	-4.04%
❌	Simulation	sort_ascii_utf8_locale	15.4 ms	16.1 ms	-4.52%
⚡	Simulation	split_bytes	401.8 µs	389 µs	+3.3%
❌	Simulation	ls_recursive_deep_tree[(200, 2)]	1.6 ms	1.7 ms	-3.47%
⚡	Simulation	df_deep_directory	385.3 µs	367.6 µs	+4.81%
⚡	Simulation	hostname_basic	159.2 µs	150.8 µs	+5.54%
⚡	Simulation	hostname_ip_lookup[100000]	117.4 µs	86.1 µs	+36.31%
⚡	Simulation	du_max_depth_balanced_tree[(6, 4, 10)]	26.1 ms	25.3 ms	+3.12%
⚡	Simulation	fold_many_lines[100000]	63.8 ms	57.3 ms	+11.41%
⚡	Simulation	fold_custom_width[50000]	25.3 ms	22.9 ms	+10.27%
⚡	Simulation	cut_fields_newline_delim	193.2 µs	185 µs	+4.44%

---

_{Comparing mattsu2020:locale_fix (205a8aa) with main (5b54e08)}

Footnotes

1. 46 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]

GNU testsuite comparison:

Skip an intermittent issue tests/tail/tail-n0f (fails in this run but passes in the 'main' branch)
Skipping an intermittent issue tests/date/date-locale-hour (passes in this run but fails in the 'main' branch)
Note: The gnu test tests/cut/bounded-memory is now being skipped but was previously passing.
Skip an intermittent issue tests/pr/bounded-memory (was skipped on 'main', now failing)

[oech3]

Is this CodSpeed change real?