Add CodSpeed performance benchmarks for linter

[codspeed-hq]

Summary

This PR sets up continuous performance tracking for Rslint using CodSpeed.

Changes

• Go benchmarks (internal/linter/linter_bench_test.go): Five benchmarks covering the linter's core execution path:

  • BenchmarkLintSingleRule_NoVar -- lints a realistic TypeScript file with the no-var rule
  • BenchmarkLintSingleRule_Eqeqeq -- lints with the strict equality rule
  • BenchmarkLintSingleRule_NoConsole -- lints with the no-console rule
  • BenchmarkLintMultipleRules -- lints with all three rules combined
  • BenchmarkLintLargeFile -- lints a 50-function TypeScript file with multiple rules to measure scaling behavior

• GitHub Actions workflow (.github/workflows/codspeed.yml): Runs benchmarks on codspeed-macro runners using walltime mode with OIDC authentication. Triggers on pushes to main, pull requests, and manual dispatch (for backtest analysis).

• CodSpeed badge added to README.md.

How it works

The benchmarks use the same program creation approach as the existing linter tests (createTestProgramWithFiles pattern). Each benchmark creates a TypeScript program from source code, then measures the time spent in RunLinterInProgram with configured rules. Setup time (program creation, temp directory) is excluded from measurements via b.ResetTimer().

Next steps

• Visit the CodSpeed dashboard after the first run to view results
• Consider adding benchmarks for type-aware rules and config parsing as the project evolves
• Performance checks on PRs will automatically flag regressions once baseline data is established

[codspeed-hq]

Congrats! CodSpeed is installed 🎉

🆕 5 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

• BenchmarkLintSingleRule_NoConsole: 1.2 ms
• BenchmarkLintLargeFile: 9.3 ms
• BenchmarkLintMultipleRules: 1.2 ms
• BenchmarkLintSingleRule_Eqeqeq: 1.2 ms
• BenchmarkLintSingleRule_NoVar: 1.2 ms

---