Hi, Could you clarify, perhaps in the README.markdown file, how to achieve the following:
1) Clone an empty repository
2) $ cd my-clone
3) Import a darcs repository history using darcs-to-git into my clone
There is one crucial step that I wish to avoid, which appears to occur automatically in darcs-to-git.. which is the 'git init' step. The repository does not need initializing as it is a clone, but it is empty - no commits exist in the clone.
Is this scenario possible with darcs-to-git ?
What's the benefit of cloning an empty repository? You're right that darcs-to-git will always create its own empty git repo, so that it can be sure that certain confusing situations arise.
That said, after importing a darcs repo, what would be the difference? It's possible to subsequently add remote upstream references etc. to the resulting git repo so that it acts the same as a cloned repo -- perhaps that's the missing piece of the puzzle?
I can think of one applied scenario.. I'd like to move code from an existing darcs repository onto GitHub. One way I might've done it:
1) Create empty repository on GitHub
2) Clone it to local machine
3) Import darcs repository to local clone using darcs-to-git
4) Using git: add everything, commit, and push back to GitHub
As darcs-to-git does not allow this to happen (importing to an existing empty repository), your suggestion of using the git remote repositories will probably work. It's just a small extra step - not a big deal, only slightly less convenient.
It should be as simple as "git remote add origin some_git_url" and then "git checkout -B -t master origin/master".
I'm going to stick with hiding the creation of the repo because there's so much scope for messed-up imports when starting with a pre-created repo.
(Oh, and probably also a "git remote update" between those two other commands.)
Sure thing, that's entirely reasonable. Thanks for responding.