Clean up test and benchmark infrastructure

[dv00d00]

Summary

• move the checked-in HTML corpus out of the misleading temp directory and update every test, benchmark, script, and documentation path
• remove a redundant test-only string helper and an unnecessary timing delay from the backpressure coverage
• apply hardware counters to every curated benchmark tier and report the actual corpus size in the tokenizer smoke runner

No corpus content or product code changes.

Validation

• dotnet build AngleSharp.ReadOnlyDom.slnx -c Release --no-restore -m:1 (0 warnings)
• 179,562/179,562 net10.0 tests passed
• tokenizer smoke: 57/57 documents, 38,831,169 UTF-8 bytes
• QQ scraper and UTF-8 DOM correctness checks passed
• BenchmarkDotNet discovered all maintained suites
• all PowerShell benchmark scripts parsed successfully