“The implication here is that any code committed to a public repository may be accessible forever as long as there is at least one fork of that repository,” the report’s authors claim.
Am I dumb or is this exactly the purpose of forks? I feel like I’m missing something.
The article is really not clear. Is it saying if a project is forked, then the original is made private, the fork can access data from the private fork?
Is this saying people misunderstand git and think committing a deletion makes people unable to access the previous version? Or is it saying the sharing between public and private repos can expose keys in private repos?
If you accidentally commit an API key into a public repository… you need to roll that key. Even if it was deleted completely, someone still could have accessed it while it was there.
from their actual report
I’m still not sure that answers it. If I fork a project, and the upstream project commits an API key (after I’ve forked it), then they delete the commit, does this commit stay available to me (unexpected behaviour)? Or is it only if I sync that commit into my repo while it’s in the upstream repo (expected behaviour)?
Or is it talking about this from a comment here:
Someone replies and said by having garbage collection kick in it removes this unconnected commit, but it’s not clear to me whether this works for github or just the local git repo.
Perhaps the issue is that these commits are synced into upstream/downstream repos when synced when they should not be?
Like I said, I’m really confused about the specifics of this.
I think Github keeps all the commits of forks in a single pool. So if someone commits a secret to one fork, that commit could be looked up in any of them, even if the one that was committed to was private/is deleted/no references exist to the commit.
The big issue is discovery. If no-one has pulled the leaky commit onto a fork, then the only way to access it is to guess the commit hash. Github makes this easier for you:
I think all GitHub should do is prune orphaned commits from the auto-suggestion list. If someone grabbed the complete commit ID then they probably grabbed the content already anyway.
Thanks, I think that explains it a bit more. It is unexpected to me, as a non-git expert, and I’m sure many others.
I guess the funny thing is that each Git commit is internally just a file. Branches and tags are just links to specific commit files and of course commits link to their parents. If a branch gets deleted or jumped back to a previous commit, the orphaned commits are still left in the filesystem. Various Git actions can trigger a garbage collection, but unless you generate huge diffs, they usually stick around for a really long time. Determining if a commit is orphaned is work that Git usually doesn’t bother doing. There’s also a reflog that can let you recover lost commits if you make a mistake.
In my experience with GitHub, dropped commits remain indefinitely accessible. I use this to my advantage on pull requests with lots of good commit context that I don’t want totally lost in a squash: by copying result of
git log --oneline main...into the PR body. The SHAs remain accessible even after I force push my branch down to a single commit.I think there is a theoretical limit to how long these commits remain accessible, but I haven’t ever hit it in my daily usage.
Ah thanks, this explains it a bit more.