Post
OK I know submodules are evil, you don't have to tell me, but #git is being a git.
I change the origin/url and so on for codeberg, but git lingers somehow.
I am finding I have to submodule update init remote recursive several time to finally get it to stick, and even have to remove and re-clone to remove last references to GitHub lurking in .git/config or somewhere.
Getting a state that you can do a clean clone --recursive is proving a pain.
Getting there slowly, 33 of 84 repositories moved.
@revk You should probably look at `git submodule sync`
@revk are you aware of subtrees?
@dysfun Yes, and I did look, and am considering.
Can you help me quite understand - if I make use of subtrees, can someone simply git clone --recursive the main repository and get all the subtrees cleanly.
I saw mention of URLs being in the repository (.gitmodules) being bad somehow - I assume somehow it knows where subtrees come from.
I should just test...
Also, my submodules ate https: as expected to be read only. I assume subtrees need to be ssh: meaning someone cloning needs a login/key.
@revk the files are copied locally into the tree, so there is no longer a submodule as such, it becomes part of the repo. so it's just git clone the "outer" repo and you're done.
@dysfun Ah, OK, but it must know the origin to allow it to be updated / pulled at subtree level still.
Sorry if stupid questions, yes, I should test.
@revk yes, you will want an update script, just like you have your submodule update script.
@dysfun Hmm, my update script is basically git submodule update --init --recursive --remote.
You mean I need a script to say to update each subtree and where it comes from?
I mean, not too bad.
I suspect I move all to codeberg before even starting a move to subtrees 🙂
