dd: stop at a zero factor in a multiplier expression

[oliver-kopcik]

Fixes #14160.

The bug

A zero factor makes a whole multiplier expression zero, so the factors after it are never looked at. parse_bytes_with_opt_multiplier parsed every factor up front instead, so a later factor that does not fit in a u64 failed the whole argument:

$ dd count=00x9999999999999999999999999999999999999999999999999999999999999 </dev/null
dd: invalid number: '00x999…': Value too large for defined data type   # exit 1

It should copy zero blocks and exit successfully, which it did before 7f9b9a6 ("dd: reject a number that does not fit in u64") turned ParseSizeError::SizeTooBig into an error rather than u64::MAX. The zero factor case was not covered there.

After this change:

$ dd count=00x9999999999999999999999999999999999999999999999999999999999999 </dev/null
0+0 records in
0+0 records out
0 bytes copied, 0.0003494 s, 0.0 B/s                                   # exit 0

The change

Return as soon as a factor parses to zero, so the remaining factors are never parsed. A number that does not fit in a u64 on its own stays an error, so the behaviour 7f9b9a6 added is kept.

The zero-multiplier warning is only a check on the literal text and needs no parsing, so it moved into its own pass over the factors. That keeps it reported once per "0" factor — count=0x0x1 still warns twice, as test_zero_multiplier_warning requires. Folding it into the short-circuiting loop would have silently dropped the second warning.

Tests

Added a unit test in parseargs.rs and an end-to-end test in tests/by-util/test_dd.rs. Both fail on main and pass with the change — verified by reverting just the if num == 0 block and re-running:

test parseargs::tests::test_parse_bytes_with_opt_multiplier_zero_factor ... FAILED
test test_dd::test_zero_factor_skips_the_rest_of_the_multiplier ... FAILED

The unit test covers a zero factor leading, in the middle, and trailing, with and without an oversized factor after it.

Run locally on Windows (x86_64-pc-windows-msvc):

• cargo test -p uu_dd --lib — 84 passed, 0 failed
• cargo test --features dd --test tests -- test_dd — 111 passed, 0 failed
• cargo fmt --all -- --check — clean
• cargo clippy -p uu_dd --all-targets — no new warnings

Note that this does not make GNU's tests/dd/misc.sh pass on its own; as #14160 says, that test then reaches a second, separate difference (failed to seek in output file: Illegal seek).

[github-actions]

GNU testsuite comparison:

Skipping an intermittent issue tests/date/date-locale-hour (passes in this run but fails in the 'main' branch)
Congrats! The gnu test tests/dd/misc is no longer failing!
Skip an intermittent issue tests/pr/bounded-memory (was skipped on 'main', now failing)