Optimizing the raw HTML post-processor

[pawamoy]

Locally, I was able to shave 20 seconds off a MkDocs build (from ~50s to ~30s) by optimizing a bit the raw HTML post processor. Would you be open to a PR that optimizes the post-processor code a bit? Any guidance on how to provide benchmarks and results to confirm to performance gain?

[waylan]

You are certainly welcome to contribute in that way. As far as benchmarks, if you can provide a reproducible command that can be run both before and after the changes which shows a clear performance boost, then that should be sufficient. Note that the command should be run directly on Markdown without any additional third-party code (such as MkDocs). Generally, Python's own built-in tools work fine.