This is a quick fix for the issue I noted on the mailing list, if a clone or fetch fails we remove the repository directory tree and start from scratch. This only occurs when the fetch call fails, and not when we kill a clone that is taking too long.
I also changed the way we call Markdown, to use their newer option for safe_mode that lets you escape html rather than just removing it. It's useful to be able to write raw html or xml in a technical blog post.
This is a quick fix for the issue I noted on the mailing list, if a clone or fetch fails we remove the repository directory tree and start from scratch. This only occurs when the fetch call fails, and not when we kill a clone that is taking too long.