fix(truncate): implement --io-blocks

[ThomasHabets]

Scale SIZE by the preferred I/O block size of each output file, using the parent filesystem for files that do not exist yet.

Before

truncate --io-blocks --size=1 file produced a 1-byte file.

After

The same command produces a 4096-byte file on a filesystem with 4096-byte I/O blocks.

There is an unrelated bug in that translate macros seem to do some rounding, probably by needlessly converting to a float somewhere. Example:

$ cargo run --  truncate --io-blocks -s 18446744073709551615 a
truncate: overflow in 18446744073709552000 * 4096 byte blocks

I'm looking into that separately. I guess this PR could choose to print the string version of the number, but that seems like a less elegant solution?

[github-actions]

GNU testsuite comparison:

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

[codspeed-hq]

Merging this PR will not alter performance

✅ 339 untouched benchmarks
⏩ 46 skipped benchmarks¹

---

_{Comparing ThomasHabets:truncate-blocks (d6e060f) with main (ef50d8b)}

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. ↩