This PR attempts to safety wrap some rough edges, by TPL-ify the existing Thread code, replacing unsafe and pointer code with Span<T> (where possible), improving performance in caching and P/Invoke usage, and fixing some NREs and race conditions found during testing.
This PR follows the working baseline.
This PR attempts to safety wrap some rough edges, by TPL-ify the existing Thread code, replacing unsafe and pointer code with
Span<T>(where possible), improving performance in caching and P/Invoke usage, and fixing some NREs and race conditions found during testing.