Add CodSpeed performance benchmarks

[codspeed-hq]

Summary

• Added performance benchmarks using pytest-codspeed to track the performance of suby's core functions: argument parsing (split_argument, convert_arguments) and end-to-end subprocess execution (run).
• Created a GitHub Actions workflow (.github/workflows/codspeed.yml) that runs benchmarks in simulation mode on pushes to main, pull requests, and manual triggers. Uses OIDC authentication.
• Added the CodSpeed badge to the README.

What's benchmarked

• split_argument: simple command strings and strings with quoted arguments
• convert_arguments: single string, multiple strings, Path objects, and split-disabled mode
• run: end-to-end subprocess execution with a minimal command

Next steps

• Once this PR is merged, CodSpeed will start tracking performance on every push to main and on pull requests.
• Visit the CodSpeed dashboard to view results and configure performance checks.
• You can expand the benchmark suite over time to cover additional scenarios (e.g., commands with cancellation tokens, custom callbacks, or timeouts).

Test plan

• Benchmarks run successfully locally with pytest --codspeed
• Benchmarks verified with codspeed run -m simulation -- pytest tests/test_benchmarks.py --codspeed
• Ruff linting passes on the new benchmark file

[codspeed-hq]

Congrats! CodSpeed is installed 🎉

🆕 7 new benchmarks were detected.

You will start to see performance impacts in the reports once the benchmarks are run from your default branch.

Detected benchmarks

• test_bench_split_argument_simple: 103.6 µs
• test_bench_split_argument_with_quotes: 115.1 µs
• test_bench_convert_arguments_single_string: 91.4 µs
• test_bench_convert_arguments_multiple_strings: 124 µs
• test_bench_convert_arguments_no_split: 21 µs
• test_bench_run_simple_command: 768.3 µs
• test_bench_convert_arguments_with_path: 79.4 µs

---

[pomponchik]

Added it in #7.