ls: honor LC_NUMERIC for the -h decimal separator

[Jorge-Polanco-Roque]

Fixes the ls part of #14232.

ls -lh (and -h) rendered the fractional size with a period regardless of LC_NUMERIC, so under a comma-decimal locale it disagreed with sort -h — breaking the common ls -h / du -h | sort -h use case.

$ LC_ALL=fr_FR.UTF-8 ls -lh 8500-byte-file   # before: 8.4K   after: 8,4K   (GNU: 8,4K)

Root cause: ls formats sizes through uucore's human_readable, which already localizes the separator via localize_decimal, but the ls crate did not enable the i18n-decimal uucore feature, so the localization compiled out. Enabling it mirrors what du already does (#12357); du and numfmt were already handled, so this closes the remaining ls gap.

Testing: added test_ls_h_locale_decimal_separator mirroring test_du_h_locale_decimal_separator. Verified on the standalone ls binary too (with the feature 8,4K, without it 8.4K). The full ls suite, cargo fmt, and clippy -D warnings pass.

[github-actions]

Binary size comparison:

Individual binary size comparison VS main (threshold: >=5% AND >=4 KB).

Total size of compared binaries: 152.18 MB (+1004 KB, +0.65%)

Significant per-binary changes:
  comm     1.12 MB ->    2.33 MB  (+1.21 MB, +107.29%)

[github-actions]

GNU testsuite comparison:

Skip an intermittent issue tests/pr/bounded-memory (fails in this run but passes in the 'main' branch)
Skip an intermittent issue tests/tail/inotify-dir-recreate (fails in this run but passes in the 'main' branch)
Skipping an intermittent issue tests/misc/io-errors (passes in this run but fails in the 'main' branch)
Skipping an intermittent issue tests/tail/tail-n0f (passes in this run but fails in the 'main' branch)
Congrats! The gnu test tests/cut/cut-huge-range is now passing!
Skip an intermittent issue tests/cut/bounded-memory (was skipped on 'main', now failing)

[Jorge-Polanco-Roque]

The failing checks look unrelated to this change. The macOS feat_os_unix job fails only on test_tail::test_follow_name_truncate3 (5312 passed, 1 failed) — a known flaky, timing-sensitive tail test — and the wasm32-wasip1 job hit a 403 Forbidden fetching packages.microsoft.com (CI infra). This PR only touches ls (adds the i18n-decimal uucore feature). Happy to rebase or re-run if that helps.

[Jorge-Polanco-Roque]

Heads-up: both red builds look unrelated to this change.

• wasm32-wasip1: the job failed during environment setup (apt-get hit a 403 from packages.microsoft.com), before cargo compiled anything — no Compiling line and no error[E...] in the log.
• macos x86_64: the only failure is test_tail::test_follow_name_truncate3, a timing-sensitive tail --follow --name test unrelated to ls. This PR's own test, test_ls_h_locale_decimal_separator, passed on that same macOS run.

Could someone re-run the two failed jobs? I ran test_follow_name_truncate3 locally 3×, passing each time, so it looks like a pre-existing flake.