fix: csplit --suffix-format huge field width panics instead of erroring

[mvanhorn]

Summary

csplit no longer panics when --suffix-format requests an enormous field width. A format like --suffix-format=%0999999999d now returns a clean error instead of aborting with a Rust panic, because the split-name formatter propagates the formatting error rather than unwrapping it.

Why this matters

Issue #12763 reports that csplit panics on a --suffix-format with a very large field width. SplitName::get used write!/format! and unwrapped the result; for a pathological width the formatter allocation/formatting failed and the unwrap turned that into a panic, crashing the utility instead of reporting a usage error the way GNU csplit does.

The formatter now returns io::Result<String> and the caller surfaces a proper csplit error, so an unusable suffix format is rejected with a message and a non-zero exit rather than a panic.

Testing

cargo test -p uu_csplit passes, including a new test_csplit case asserting that a huge --suffix-format width errors out (non-zero exit, no panic) instead of crashing. cargo clippy -p uu_csplit is clean.

Fixes #12763

[codspeed-hq]

Merging this PR will not alter performance

✅ 331 untouched benchmarks
⏩ 46 skipped benchmarks¹

---

_{Comparing mvanhorn:fix/12763-csplit-suffix-format-width-panic (24b9c6a) with main (dad997e)}

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/cut/bounded-memory (fails in this run but passes in the 'main' branch)
Skip an intermittent issue tests/rm/isatty (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)
Skipping an intermittent issue tests/tail/symlink (passes in this run but fails in the 'main' branch)
Note: The gnu test tests/rm/many-dir-entries-vs-OOM is now being skipped but was previously passing.
Congrats! The gnu test tests/seq/seq-epipe is now passing!

[sylvestre]

sorry but could you please rebase it ? thanks

[mvanhorn]

Rebased onto main. There were a couple of real conflicts from the csplit error-type refactor (the new strip_errno IoError with #[from]): kept your strip_errno formatting but dropped #[from], since this fix needs a custom From<io::Error> that maps OutOfMemory to a MemoryExhausted error (that's what turns a huge --suffix-format width into a clean error instead of a panic). cargo build, cargo test -p uu_csplit (37 pass), fmt and clippy all clean.