blob: e81266a7d446648ac280b3afa309efec6a5c88b4 (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
|
[[!comment format=mdwn
username="CandyAngel"
subject="comment 2"
date="2015-07-15T10:20:55Z"
content="""
Been thinking about this as I am getting close to needing it, but would like some advice.
My current plan is to copy the symlink to the target annex, add it to the index (fix it?), copy the source file from $source/.git/annex/objects to $target/.git/annex/tmp, then use 'reinject $target/.git/annex/tmp/$keyed_file $path_to_symlink'.
As far as I can tell, this is safest way (uses mostly git-annex) to transfer a file between annexes. However, when transferring a directory of files, this will end up with 1 commit per file on the git-annex branch, which may be a problem.
Is there any easy way to make this \"atomic\", so that git-annex will only get a commit if everything went okay and if not, revert any changes to $target? Am I looking at 'git stash', recording the master/git-annex references before the move and resetting to them in case of an error or rebasing(fixup) git-annex on success?
"""]]
|
