split: preserve non-UTF-8 bytes in output filename generation

[can1357]

uutils split accepts non-UTF-8 prefix and suffix inputs but converts them with to_string_lossy() when building chunk filenames. GNU keeps pathname bytes intact, while uutils rewrites invalid bytes to UTF-8 replacement characters.

Reproduction Steps

d=$(mktemp -d); p=$(printf "p\377"); printf "AB" | split -b1 - "$d/$p"; ls "$d" | od -An -tx1
# Expected (GNU): 70 ff 61 61 0a 70 ff 61 62 0a
# Actual (uutils): 70 ef bf bd 61 61 0a 70 ef bf bd 61 62 0a

Impact

Chunk files are created under rewritten names instead of the requested byte paths. This breaks GNU compatibility in non-UTF-8 environments and can cause filename collisions or misdirected output files.

[codspeed-hq]

Merging this PR will improve performance by 6.79%

⚠️ 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

⚡ 1 improved benchmark
✅ 297 untouched benchmarks
⏩ 48 skipped benchmarks¹

Performance Changes

	Mode	Benchmark	BASE	HEAD	Efficiency
⚡	Simulation	split_bytes	429.8 µs	402.5 µs	+6.79%

---

_{Comparing can1357:split-preserve-non-utf8-bytes-in-output-filename-generation (d5421ab) with main (ef8e45c)}

Footnotes

1. 48 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. ↩