One thing here is that I'm parsing an abbreviated oid out of the patch file into the id's in the delta with no warning about the fact that they were abbreviated or what the length of known bytes was.

I think the right thing to do here is add a flag that indicates it's abbreviated and the number of bytes known; this seems a bit superior to always setting the full id length, since most deltas are not created from patch files and will always have a full id. Only the bits that work with patches and their deltas need to know whether there is an abbreviation and the rest of diff, merge, etc can continue blissfully treating the id as complete.

Other opinions, though, are quite welcome.

And also a big thanks to @arrbee for making git_patch so sane that I could mostly plug right into it. I had this populating a different structure, first, as a proof of concept before throwing that away and moving it into the existing patch. Very glad that worked out the way it did; though I need to give this a second pass and make sure that I'm not neglecting to populate any fields. But the tests to apply from the git_patch make me pretty confident.

/cc @brianmario - I think we were just talking about stuff related to this

What's the status of this?

Totally stalled and not something that I'm going to have time to attack in the near future.

Can I take on this PR and try and hack on it?

This is so old that we targeted the development branch. Closing this, moving this to master.