How do I clone over HTTP a repository that has no info/refs?

Question

Given a repository served over HTTP whose owner forgot to chmod +x hooks/post-update, is there a workaround for cloning it?

I tried running wget --mirror url, but rather than fetching the subtree only, it tried to mirror the entire site—which I assume happened due to the parent-directory links in the autogenerated index.html resources.

Answer 1

If you just want the source code to build and don't really need the repo, try adding --no-parent to wget. That should limit wget to the subtree.

Apart from that, email the person who runs the repository and ask them to fix it.