Closed l1x closed 9 months ago
The hash is how we do atomic updates, and how we know what we think we have already (in case of a crash or a long-lived volume). When upstream changes, we sync, prepare a new worktree, then flip the link.
Otherwise your client might see a tree that is half updated in the middle of a pull.
Instead of accessing it via the hash directory, you can use the symlink, which should always point to the latest hash).
Thanks Tim!
I have to following setup:
For some reason this results in a data/9cb27e1a1eb90c2832c22447c5a88085f132e62d/..... The last commit hash is included in the path.
Is there a way to influence this behaviour so that the commit hash is not showing up in the folder?