I have a 3 projects I'd like to fork. They're all related to each other - changing one will likely require a change to another. Because they're all related, I'd like to create 1 repository for the forks, while maintaining the ability to pull down updates from each original.
How would I setup my git repository?
These are preliminary thoughts so I wouldn't be surprised if this is crazy/stupid. Is it?
You can always use "git remote add <name> <url>
" to add more repositories as source.
You may be interested in git-submodule functionality. Quote from here:
Git's submodule support allows a
repository to contain, as a
subdirectory, a checkout of an
external project. Submodules maintain
their own identity; the submodule
support just stores the submodule
repository location and commit ID, so
other developers who clone the
containing project ("superproject")
can easily clone all the submodules at
the same revision. Partial checkouts
of the superproject are possible: you
can tell Git to clone none, some or
all of the submodules.
There's one possibility I've never tried "for real" but that might work well.
Let's say you have three projects creatively named Proj1, Proj2 and (guess it) Proj3. Let's also say that all or some of them depend on external libraries, for example one called Lib1.dll and another called SuperLib.dll.
First, for the folder structure, I'd do something like:
MyCode\
build_all.bat
Docs\
[...]
Libs\
Lib1.dll
SuperLib.dll
Proj1\
[...]
Proj2\
[...]
Proj3\
[...]
Then you would create one repository on each of the ProjX folders:
cd \MyCode\Proj1
git init
cd \MyCode\Proj2
git init
cd \MyCode\Proj3
git init
Then you would create a repository for the whole thing:
cd \MyCode
git init
The contents of this repository, for git, would be the Libs
folder and its contents, the build_all.bat
build script, and 3 subprojects.
So once you change your source code, let's say Proj1\Main.cs
, the whole thing wouldn't see the modification. You would have to git commit
on Proj1, and then you would go to the whole thing and then git commit
it (now git will be able to see that Proj1 has been modified).
Using this method, the history of the whole thing would simply state when the ProjX were modified, but it would not directly keep track of individual files. Howver, as expected, the repository on ProjX would keep track of every file, as usual.
If you give it a try on a "real project", let me know the results!
Good luck!
I think I should be using the subtree merge strategy. I'll have to actually try it out and see if it works well. I'll mark this answer as accepted if this turns out to be a good approach.
In the mean time, I'm still open for suggestions.