git子模块的common子模块

Question

I have an application main which consists of multiple submodules A , B and utility_functions : Both of these submodules A and B by itself include the submodule utility_functions .

main
|
|---- A
|     |
|     |---- utility_functions
|
|---- B
|     |
|     |---- utility_functions
|
|---- utility_functions

Right now I have the utility_functions three times in my project. Usually sub_A , sub_B and main should all point to the same commit of utility_functions .

It is necessary that the code from the repository A is run-able by itself. If I only clone A it should include my utility_functions

A
|
|---- utility_functions

Is there a way to only tell my submodules included in the main repository to rely on the same utility_functions ? Or does this break the idea of submodules because they can point to different commits? Or is this simply a bad idea?

main
|
|---- A
|     |
|     |----
|         |
|---- B   |
|     |   |
|     |   |
|     |   |
|---- utility_functions

Answer 1

Or does this break the idea of submodules because they can point to different commits?

Yes, that wouldn't be practical precisely because of that reason.

If you are sure that A and B should always reference the same version of utility_functions , then clone your main repo and update only A , B and utility_functions (without the recursive option), and make a symlink from A/utility_functions and B/utility_functions to ../utility_functions .