git LFS support

[philocalyst]

This Pull Request fixes #2812

It changes the following:

• This adds Git LFS support through the git-lfs implementation
  • There ARE some failing test-cases within the implementation, it's not as perfect as shelling out. But it DOES happen to work!
  • Most of the fails are related to bulletproofing, like ssh, credentials, etc. So... kind of yikes, but I know this will improve. (And we don't have another option)
• This is a very basic/introductory PR, laying the groundwork for the finer features like lock/unlock, pruning, status, etc.
• I've elected to go with a simple [LFS] marker rather than some new symbol.
• • Respects "GIT_LFS_SKIP_SMUDGE"
• Tries to be pretty careful with attempting smudge and related LFS operations
• In retrospect perhaps a bit log-heavy compared to what I see in related code, part of that was debugging, so let me know if it's overwhelming.
• Went with git2 in the asyncgit portion of this PR because of the embedded nature of a lot of these new functions.

I followed the checklist:

• I added unittests
• I ran make check without errors
• I tested the overall application
• I added an appropriate item to the changelog

[philocalyst]

Converting to draft because everytime I try to add new features I go up against the brutal limitations of the library. It's the only one available, but I find myself reimplementing transport mechanisms and that's where it's too messy for me.

Are we okay with shelling out to the git-lfs binary for those parts? And keeping it mixed? Think that's the most pragmatic in the short-term.