This is a newb question. I don't know whether this is a bug or the way git-annex is intended to function.
I have two annex repos connected to each other. My idea was to have the first repository add files, which would then be moved to the second repository for storage. After moving, repo1 would be empty again, empty and clean of any symlinks.
But after I 'git-annex move * --to repo2' broken symlinks remain in repo1. I don't want any broken/unused symlinks to remain in repo1 for object data it doesn't currently have (even if those files remain in the repository itself).
Is there a way I can clean/remove broken symlinks to object data when those objects aren't present, so the directory only contains symlinks when the repo currently has the object data for those files?
git-annex is behaving as expected here. The broken symlink allows you to run
git annex geton it to get the file content back into the repository, orgit mvto rename the file even though its content is not present, etc.You can probably accomplish what you want by using git branches. You want a branch for repo2 that has all the files, and a branch for repo1 that has only the files in repo1. git-annex doesn't maintain such branches for you, but you can probably come up with a way to create such branches (
git annex findwill be useful when scripting up a solution).Or you can adopt the approach the git-annex assistant uses for archived files -- a archive/ directory, where files are moved when they're no longer wanted in the local system, so that their symlinks don't clutter up the view, while still being easily accessible when the time comes to pull something out of the archive.