Write documentation for LFS
The new LFS (large file support) feature is nice in theory, but it does absolutely nothing on my computer and server. I have installed the client 1.1.1 for Windows and configured it as described in all the three-line documentations available everywhere. All commits and pushes take their usual path, there's nothing to be seen of anything LFS. The server is version 8.4.something.
The Git LFS project itself doesn't have any usable documentation beyond boring manpages at all. The GitLab project (http://doc.gitlab.com/ce/workflow/lfs/manage_large_binaries_with_git_lfs.html) has not much more on top of that.
I need information about the following topics:
- What exactly does LFS do? How does it do that, what mechanisms does it use, and when does it come into action?
- What path does a committed file under LFS control take? Is it created as a regular Git object under .git/ or is it stored elsewhere? How is it handled on the next and all subsequent pushes to different remotes?
- What happens if a remote doesn't support LFS, say a network or USB drive?
- How much does vanilla Git know about LFS, how far can it regard this extension and what does or doesn't work if the LFS client is not installed?
- Does it work with an HTTP-only (not HTTPS) server, and how? Can it support SSH for everything as well to prevent the need to enter a rarely used web access password for every single push and fetch now? (That makes using SSH and keys pointless, you might as well use HTTP(S) for every remote Git operation. Unfortunately, Git clients won't remember username+password like SVN clients do.)
- How can things be debugged if nothing works? Should I throw my computer away and restart, or are there other analysis possibilities?
Obviously, I can't contribute on this one because I have no idea about the contents to be written.