On 18/12/2013 04:33, Marvin Humphrey wrote:
On Tue, Dec 17, 2013 at 3:15 AM, Nick Wellnhofer <[email protected]> wrote:
It will be hard to track changes across these renames, so I think we could
as well start the Clownfish repo from scratch.
Works for me.
I didn't think about the Clownfish compiler. A better approach is to use
"git filter-branch --subdirectory-filter". I followed the instructions here
http://stackoverflow.com/questions/359424/detach-subdirectory-into-separate-git-repository
and ran
$ cd ..
$ git clone lucy clownfish
$ cd clownfish
$ git remote rm origin
$ git filter-branch --tag-name-filter cat --prune-empty
--subdirectory-filter clownfish HEAD
$ git reset --hard
$ git for-each-ref --format="%(refname)" refs/original/ | xargs -n 1 git
update-ref -d
$ git reflog expire --expire=now --all
$ git gc --aggressive --prune=now
This will lose most of the Clownfish runtime history but keeps the
history of CFC completely intact. The resulting repo is 8.8M on my machine.
I also tried the "new way" using "git subtree":
$ git subtree split -P clownfish -b clownfish-only
$ cd ..
$ mkdir clownfish
$ cd clownfish
$ git init
$ git pull ../lucy clownfish-only
But this resulted in a lot of useless history.
The same approach should work for Charmonizer.
I'm fine with naming the Clownfish repository lucy-clownfish.git, even though
the product will be "Apache Clownfish" and not "Apache Lucy Clownfish". We'll
cross the TLP bridge when we come to it.
Likewise with the Charmonizer repo being named lucy-charmonizer.git. With
Charmonizer, there are not even any plans as of yet to release the full
product independently.
+1
I think we should call a vote on the repo split first.
Nick
