grep: expand bundled numeric context options

[ColumbusLabs]

Summary

• expand numeric context shorthands embedded in short-option clusters
• preserve standalone multi-digit context values and left-to-right option overrides
• leave attached and separate values for value-taking options, long options, operands, post--- arguments, and non-UTF-8 arguments unchanged
• cover the reported fuzzer case and parser non-regressions

The implementation is based only on the behavior specified in issue #68 and the original uutils grep code. No GNU source code was consulted or used.

Verification

• focused numeric shorthand tests: 5 of 5 passed
• full test suite: 11 unit and 94 integration tests passed
• cargo clippy --all-targets --all-features -- -D warnings
• cargo fmt --all -- --check
• git diff --check

Closes #68

[codspeed-hq]

Merging this PR will not alter performance

✅ 10 untouched benchmarks
⏩ 17 skipped benchmarks¹

---

_{Comparing ColumbusLabs:agent/bundled-num-shorthand (56b9aab) with main (aa7f3c9)}

Footnotes

1. 17 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. ↩

[lhecker]

I'm not entirely sure if I agree with the way this PR is written. It removes the Vec collect() optimization (os_args() returns a Vec IntoIterator), adds full UTF8 validation (is that necessary?), and is overall a lot more verbose than the old code.

I haven't thought it through, and so I may be mistaken, but I intuitively feel like the issue should be moderately easy to solve within the old code's structure.